Journal Articles
2020 |
Zdravevski, Eftim; Lameski, Petre; Apanowicz, Cas; Ślȩzak, Dominik From Big Data to business analytics: The case study of churn prediction Journal Article In: Applied Soft Computing, vol. 90, pp. 106164, 2020, ISSN: 1568-4946. @article{ZDRAVEVSKI2020106164, The success of companies hugely depends on how well they can analyze the available data and extract meaningful knowledge. The Extract-Transform-Load (ETL) process is instrumental in accomplishing these goals, but requires significant effort, especially for Big Data. Previous works have failed to formalize, integrate, and evaluate the ETL process for Big Data problems in a scalable and cost-effective way. In this paper, we propose a cloud-based ETL framework for data fusion and aggregation from a variety of sources. Next, we define three scenarios regarding data aggregation during ETL: (i) ETL with no aggregation; (ii) aggregation based on predefined columns or time intervals; and (iii) aggregation within single user sessions spanning over arbitrary time intervals. The third scenario is very valuable in the context of feature engineering, making it possible to define features as “the time since the last occurrence of event X”. The scalability was evaluated on Amazon AWS Hadoop clusters by processing user logs collected with Kinesis streams with datasets ranging from 30 GB to 2.6 TB. The business value of the architecture was demonstrated with applications in churn prediction, service-outage prediction, fraud detection, and more generally — decision support and recommendation systems. In the churn prediction case, we showed that over 98% of churners could be detected, while identifying the individual reason. This allowed support and sales teams to perform targeted retention measures. |
Ferreira, José M; Pires, Ivan Miguel; Marques, Gonçalo; Garcia, Nuno M; Zdravevski, Eftim; Lameski, Petre; Flórez-Revuelta, Francisco; Spinsante, Susanna Identification of Daily Activites and Environments Based on the AdaBoost Method Using Mobile Device Data: A Systematic Review Journal Article In: Electronics, vol. 9, no. 1, pp. 192, 2020. @article{ferreira2020identification, |
Ferreira, José M; Pires, Ivan Miguel; Marques, Gonçalo; Garcia, Nuno M; Zdravevski, Eftim; Lameski, Petre; Flórez-Revuelta, Francisco; Spinsante, Susanna; Xu, Lina Activities of Daily Living and Environment Recognition Using Mobile Devices: A Comparative Study Journal Article In: Electronics, vol. 9, no. 1, pp. 180, 2020. @article{ferreira2020activities, |
Villasana, María Vanessa; Pires, Ivan Miguel; Sá, Juliana; Garcia, Nuno M; Zdravevski, Eftim; Chorbev, Ivan; Lameski, Petre; Flórez-Revuelta, Francisco Promotion of Healthy Nutrition and Physical Activity Lifestyles for Teenagers: A Systematic Literature Review of The Current Methodologies Journal Article In: Journal of Personalized Medicine, vol. 10, no. 1, 2020, ISSN: 2075-4426. @article{jpm10010012, Amid obesity problems in the young population and apparent trends of spending a significant amount of time in a stationary position, promoting healthy nutrition and physical activities to teenagers is becoming increasingly important. It can rely on different methodologies, including a paper diary and mobile applications. However, the widespread use of mobile applications by teenagers suggests that they could be a more suitable tool for this purpose. This paper reviews the methodologies for promoting physical activities to healthy teenagers explored in different studies, excluding the analysis of different diseases. We found only nine studies working with teenagers and mobile applications to promote active lifestyles, including the focus on nutrition and physical activity. Studies report using different techniques to captivate the teenagers, including questionnaires and gamification techniques. We identified the common features used in different studies, which are: paper diary, diet diary, exercise diary, notifications, diet plan, physical activity registration, gamification, smoking cessation, pictures, game, and SMS, among others. |
Faria, Rui; Lopes, Inês; Pires, Ivan Miguel; Marques, Gonçalo; Fernandes, Solange; Garcia, Nuno M; Lucas, José; Jevremović, Aleksandar; Zdravevski, Eftim; Trajkovik, Vladimir Circular Economy for Clothes Using Web and Mobile Technologies—A Systematic Review and a Taxonomy Proposal Journal Article In: Information, vol. 11, no. 3, 2020, ISSN: 2078-2489. @article{info11030161, Nowadays, it is possible to buy clothing using online platforms, either by accessing online brand stores, general online stores or circular economy platforms. This paper presents a study on mobile applications that support online commerce for clothing, focusing on the review of the mobile applications with features that are characteristic of the circular economy paradigm. Findings include the fact that almost all the mobile applications analysed have pictures illustrative of the clothes and accessories that are available for trading as well as their brief description. Furthermore, this paper presents a study of various scientific articles about the circular economy of clothes and how it can be beneficial to the future of the environment. It is a junction with a Web platform for its growth and its disclosure. The paper builds conclusions upon the assumption that the circular economy is a growing business that is part of a sustainable development where the main goal is to reduce the environmental impact. The paper proposes the analysis of an innovative taxonomy of mobile applications about the circular economy. |
Pires, Ivan Miguel; Marques, Gonçalo; Garcia, Nuno M; Flórez-Revuelta, Francisco; Teixeira, Maria Canavarro; Zdravevski, Eftim; Spinsante, Susanna; Coimbra, Miguel Pattern Recognition Techniques for the Identification of Activities of Daily Living Using a Mobile Device Accelerometer Journal Article In: Electronics, vol. 9, no. 3, 2020, ISSN: 2079-9292. @article{electronics9030509, The application of pattern recognition techniques to data collected from accelerometers available in off-the-shelf devices, such as smartphones, allows for the automatic recognition of activities of daily living (ADLs). This data can be used later to create systems that monitor the behaviors of their users. The main contribution of this paper is to use artificial neural networks (ANN) for the recognition of ADLs with the data acquired from the sensors available in mobile devices. Firstly, before ANN training, the mobile device is used for data collection. After training, mobile devices are used to apply an ANN previously trained for the ADLs’ identification on a less restrictive computational platform. The motivation is to verify whether the overfitting problem can be solved using only the accelerometer data, which also requires less computational resources and reduces the energy expenditure of the mobile device when compared with the use of multiple sensors. This paper presents a method based on ANN for the recognition of a defined set of ADLs. It provides a comparative study of different implementations of ANN to choose the most appropriate method for ADLs identification. The results show the accuracy of 85.89% using deep neural networks (DNN). |
Ponciano, Vasco; Pires, Ivan Miguel; Ribeiro, Fernando Reinaldo; Marques, Gonçalo; Garcia, Nuno M; Pombo, Nuno; Spinsante, Susanna; Zdravevski, Eftim Is The Timed-Up and Go Test Feasible in Mobile Devices? A Systematic Review Journal Article In: Electronics, vol. 9, no. 3, 2020, ISSN: 2079-9292. @article{electronics9030528, The number of older adults is increasing worldwide, and it is expected that by 2050 over 2 billion individuals will be more than 60 years old. Older adults are exposed to numerous pathological problems such as Parkinson’s disease, amyotrophic lateral sclerosis, post-stroke, and orthopedic disturbances. Several physiotherapy methods that involve measurement of movements, such as the Timed-Up and Go test, can be done to support efficient and effective evaluation of pathological symptoms and promotion of health and well-being. In this systematic review, the authors aim to determine how the inertial sensors embedded in mobile devices are employed for the measurement of the different parameters involved in the Timed-Up and Go test. The main contribution of this paper consists of the identification of the different studies that utilize the sensors available in mobile devices for the measurement of the results of the Timed-Up and Go test. The results show that mobile devices embedded motion sensors can be used for these types of studies and the most commonly used sensors are the magnetometer, accelerometer, and gyroscope available in off-the-shelf smartphones. The features analyzed in this paper are categorized as quantitative, quantitative + statistic, dynamic balance, gait properties, state transitions, and raw statistics. These features utilize the accelerometer and gyroscope sensors and facilitate recognition of daily activities, accidents such as falling, some diseases, as well as the measurement of the subject’s performance during the test execution. |
Corizzo, Roberto; Ceci, Michelangelo; Zdravevski, Eftim; Japkowicz, Nathalie Scalable Auto-Encoders for Gravitational Waves Detection from Time Series Data Journal Article In: Expert Systems with Applications, pp. 113378, 2020, ISSN: 0957-4174. @article{CORIZZO2020113378, Gravitational waves represent a new opportunity to study and interpret phenomena from the universe. In order to efficiently detect and analyze them, advanced and automatic signal processing and machine learning techniques could help to support standard tools and techniques. Another challenge relates to the large volume of data collected by the detectors on a daily basis, which creates a gap between the amount of data generated and effectively analyzed. In this paper, we propose two approaches involving deep auto-encoder models to analyze time series collected from Gravitational Waves detectors and provide a classification label (noise or real signal). The purpose is to discard noisy time series accurately and identify time series that potentially contain a real phenomenon. Experiments carried out on three datasets show that the proposed approaches implemented using the Apache Spark framework, represent a valuable machine learning tool for astrophysical analysis, offering competitive accuracy and scalability performances with respect to state-of-the-art methods. |
Maresova, P; Krejcar, O; Barakovic, S; Husic, J Barakovic; Lameski, P; Zdravevski, E; Chorbev, I; Trajkovik, V Health–Related ICT Solutions of Smart Environments for Elderly–Systematic Review Journal Article In: IEEE Access, vol. 8, pp. 54574-54600, 2020, ISSN: 2169-3536. @article{Maresova2020ict, By improving the quality of life and extending the length of life, Western society is becoming an increasingly ageing population with a higher proportion of seniors. From another point of view, there is a critical shortage of care staff, both in hospitals and for in-home care. Thanks to new technology trends such as Smart Homes and Smart Furniture, there is an opportunity for increased support for seniors by utilizing new technologies. This paper presents the current trends and possibilities in applying smart information and communications technology (ICT) solutions for in-home care concerning diseases in old age. The paper consists of a systematic review according to the PRISMA methodology of the available literature in Web of Science, IEEE Xplore, PubMed, Springer, and the Espacenet patent database. Publications report the usage of some types of artificial intelligence and their implementation and non-intrusive sensing technologies. The patents review identified solutions with a focus on monitoring the state of older adults and mobility improvement. Existing ICT smart solutions must address the following issues: (1) ease-of-use; (2) invisibility and disuse that isolate older adults; (3) privacy and security; (4) affordability of technology in terms of cost; and (5) supporting elderly individuals to stay in their homes or move in different environments independently. There is a significant gap between a large number of scientific publications and commercial solutions. The existing products reflect the specifics of the diseases in a rather wider context instead of the fulfilment of exact needs. It is often stated that such devices can be used across diseases, but the direct connection and benefits for the disease is still rather weak. The challenge remains to tap the existing potential of a large number of innovative ideas on the market and improve the quality of life. |
2019 |
Loncar-Turukalo, Tatjana; Zdravevski, Eftim; Silva, José Machado; Chouvarda, Ioanna; Trajkovik, Vladimir Literature on Wearable Technology for Connected Health: Scoping Review of Research Trends, Advances, and Barriers Journal Article In: J Med Internet Res, vol. 21, no. 9, pp. e14017, 2019, ISSN: 1438-8871. @article{turukalo2019jmir, Background: Wearable sensing and information and communication technologies are key enablers driving the transformation of health care delivery toward a new model of connected health (CH) care. The advances in wearable technologies in the last decade are evidenced in a plethora of original articles, patent documentation, and focused systematic reviews. Although technological innovations continuously respond to emerging challenges and technology availability further supports the evolution of CH solutions, the widespread adoption of wearables remains hindered. Objective: This study aimed to scope the scientific literature in the field of pervasive wearable health monitoring in the time interval from January 2010 to February 2019 with respect to four important pillars: technology, safety and security, prescriptive insight, and user-related concerns. The purpose of this study was multifold: identification of (1) trends and milestones that have driven research in wearable technology in the last decade, (2) concerns and barriers from technology and user perspective, and (3) trends in the research literature addressing these issues. Methods: This study followed the scoping review methodology to identify and process the available literature. As the scope surpasses the possibilities of manual search, we relied on the natural language processing tool kit to ensure an efficient and exhaustive search of the literature corpus in three large digital libraries: Institute of Electrical and Electronics Engineers, PubMed, and Springer. The search was based on the keywords and properties to be found in articles using the search engines of the digital libraries. Results: The annual number of publications in all segments of research on wearable technology shows an increasing trend from 2010 to February 2019. The technology-related topics dominated in the number of contributions, followed by research on information delivery, safety, and security, whereas user-related concerns were the topic least addressed. The literature corpus evidences milestones in sensor technology (miniaturization and placement), communication architectures and fifth generation (5G) cellular network technology, data analytics, and evolution of cloud and edge computing architectures. The research lag in battery technology makes energy efficiency a relevant consideration in the design of both sensors and network architectures with computational offloading. The most addressed user-related concerns were (technology) acceptance and privacy, whereas research gaps indicate that more efforts should be invested into formalizing clear use cases with timely and valuable feedback and prescriptive recommendations. Conclusions: This study confirms that applications of wearable technology in the CH domain are becoming mature and established as a scientific domain. The current research should bring progress to sustainable delivery of valuable recommendations, enforcement of privacy by design, energy-efficient pervasive sensing, seamless monitoring, and low-latency 5G communications. To complement technology achievements, future work involving all stakeholders providing research evidence on improved care pathways and cost-effectiveness of the CH model is needed. |
Hoang, Yen; Pfeil, Juliane; Zagorščak, Maja; Thieffry, Axel; Zdravevski, Eftim; Ramšak, Živa; Lameski, Petre; Schulze, Sabrina; Papakonstantinou, Eleni; Papageorgiou, Louis; Singh, Tarry; Duarte-López, Ariel; Pérez-Casany, Marta Report on the “Advanced Big Data Training School for Life Sciences”, Barcelona 3th-7th September 2018 Journal Article In: EMBnet.journal, vol. 24, no. 0, pp. 917, 2019, ISSN: 2226-6089. @article{hoang2019bigdata, The “Advanced Big Data Training School for Life Sciences” took place during September 3-7, 2018, organized by the Data Management Group (DAMA-UPC) at the Technical University of Catalonia (UPC) in Barcelona, Spain. It is the follow-up training school of the first “Big Data Training School for Life Sciences”, held in Uppsala, Sweden, in September 2017, which was defined and structured at the “Think Tank Hackathon”, held in Ljubljana, Slovenia, in February 2018. The aim of this training school was to get participants acquainted with emerging Big Data processing techniques in the field of Computational Biology and Bioinformatics.This article explains in detail the development of the training school, the covered contents and the interaction of the participants within and out of the training event by the student, organizer and lecturer perspective. |
2018 |
Maresova, Petra; Tomsone, Signe; Lameski, Petre; Madureira, Joana; Mendes, Ana; Zdravevski, Eftim; Chorbev, Ivan; Trajkovik, Vladimir; Ellen, Moriah; Rodile, Kasper Technological Solutions for Older People with Alzheimer's Disease Journal Article In: Current Alzheimer Research, vol. 15, no. 10, pp. 975–983, 2018. @article{maresova2018technological, |
Pires, Ivan Miguel; Teixeira, Maria Canavarro; Pombo, Nuno; Garcia, Nuno M; Flórez-Revuelta, Francisco; Spinsante, Susanna; Goleva, Rossitza; Zdravevski, Eftim; others, Android Library for Recognition of Activities of Daily Living: Implementation Considerations, Challenges, and Solutions Journal Article In: 2018. @article{pires2018android, |
Pfeil, Juliane; Schulze, Sabrina; Zdravevski, Eftim; Hoang, Yen Report on the “Big Data Training School for Life Sciences”, 18-22 September 2017, Uppsala, Sweden Journal Article In: EMBnet.journal, vol. 23, no. 0, pp. 905, 2018, ISSN: 2226-6089. @article{pfeil2018bigdata, In September 2017 a "Big Data Training School for Life Sciences" took place in Uppsala, Sweden, jointly organised by EMBnet and the COST Action CHARME (Harmonising standardisation strategies to increase efficiency and competitiveness of European life-science research - CA15100). The week programme was divided into hands-on sessions and lectures. In both cases, insights into dealing with big amounts of data were given. This paper describes our personal experience as students’ by providing also some suggestions that we hope can help the organisers as well as other trainers to further increase the efficiency of such intensive courses for students with diverse backgrounds. |
Schulze, Sabrina; Ramšak, Živa; Hoang, Yen; Zdravevski, Eftim; Pfeil, Juliane; Duarte-López, Ariel; Baier, Uwe; Zagorščak, Maja Proceedings of the “Think Tank Hackathon’’, Big Data Training School for Life Sciences Follow-up, Ljubljana 6th – 7th February 2018 Journal Article In: EMBnet.journal, vol. 24, no. 0, pp. 912, 2018, ISSN: 2226-6089. @article{schulze2018bigdata, On 6th and 7th February 2018, a Think Tank took place in Ljubljana, Slovenia. It was a follow-up of the “Big Data Training School for Life Sciences” held in Uppsala, Sweden, in September 2017. The focus was on identifying topics of interest and optimising the programme for a forthcoming “Advanced” Big Data Training School for Life Science, that we hope is again supported by the COST Action CHARME (Harmonising standardisation strategies to increase efficiency and competitiveness of European life-science research - CA15110). The Think Tank aimed to go into details of several topics that were - to a degree - covered by the former training school. Likewise, discussions embraced the recent experience of the attendees in light of the new knowledge obtained by the first edition of the training school and how it comes from the perspective of their current and upcoming work. The 2018 training school should strive for and further facilitate optimised applications of Big Data technologies in life sciences. The attendees of this hackathon entirely organised this workshop. |
Stojanovski, Aleksandar; Zdravevski, Eftim; Koceski, Saso; Trajkovik, Vladimir Real-time sleep apnea detection with one-channel ECG based on edge computing paradigm Journal Article In: 2018. @article{stojanovski2018real, |
Pires, Ivan Miguel; Santos, Rui; Pombo, Nuno; Garcia, Nuno M; Flórez-Revuelta, Francisco; Spinsante, Susanna; Goleva, Rossitza; Zdravevski, Eftim Recognition of Activities of Daily Living Based on Environmental Analyses Using Audio Fingerprinting Techniques: A Systematic Review Journal Article In: Sensors, vol. 18, no. 1, ARTICLE NUMBER = 160, 2018, ISSN: 1424-8220. @article{pires2018sensors, An increase in the accuracy of identification of Activities of Daily Living (ADL) is very important for different goals of Enhanced Living Environments and for Ambient Assisted Living (AAL) tasks. This increase may be achieved through identification of the surrounding environment. Although this is usually used to identify the location, ADL recognition can be improved with the identification of the sound in that particular environment. This paper reviews audio fingerprinting techniques that can be used with the acoustic data acquired from mobile devices. A comprehensive literature search was conducted in order to identify relevant English language works aimed at the identification of the environment of ADLs using data acquired with mobile devices, published between 2002 and 2017. In total, 40 studies were analyzed and selected from 115 citations. The results highlight several audio fingerprinting techniques, including Modified discrete cosine transform (MDCT), Mel-frequency cepstrum coefficients (MFCC), Principal Component Analysis (PCA), Fast Fourier Transform (FFT), Gaussian mixture models (GMM), likelihood estimation, logarithmic moduled complex lapped transform (LMCLT), support vector machine (SVM), constant Q transform (CQT), symmetric pairwise boosting (SPB), Philips robust hash (PRH), linear discriminant analysis (LDA) and discrete cosine transform (DCT). |
2017 |
Starič, Kristina Drusany; Bukovec, Petra; Jakopič, Katja; Zdravevski, Eftim; Trajkovik, Vladimir; Lukanović, Adolf Can we predict obstetric anal sphincter injury? Journal Article In: European Journal of Obstetrics & Gynecology and Reproductive Biology, vol. 210, pp. 196–200, 2017, ISSN: 03012115. @article{tina2017oasi, Objective The aim of the study was to identify primiparous pregnant women with a higher risk for obstetric anal sphincter injuries (OASIS) based on obstetric characteristics (risk factors). Study design In the retrospective case control study primiparous women were examined using endoanal ultrasonography (EUS) for OASIS identification 6–12 weeks after delivery. Obstetric characteristics for OASIS were collected from the mothers' medical records. The univariate analysis of maternal (age at delivery, maternal height, weight, BMI), infant (length, weight and head circumference) and birth (pregnancy duration, labour and delivery duration, episiotomy, vacuum extraction and oxytocin augmentation) risk factors, Pearson correlations and information gain were carried out. The cut-off values for the aforementioned risk factors divided the patients into groups with higher and lower risk of OASIS. Results The data of 84 primiparous women with OASIS, and 58 without, were analysed. Those newborns born to women in the OASIS group were heavier (P textless 0.05), with the cut-off at 3420 g (72% probability of OASIS), had a larger head circumference (P textless 0.001), cut-off at 36 cm (84% probability of OASIS), and were longer (P textless 0.05), cut-off at 50.5 cm (74% probability of OASIS). The maternal age and body mass index (BMI) were risk factors for OASIS (P textless 0.05 and P textless 0.05, respectively) with a probability of 83% in women younger than 27.5 years and a 78% probability if BMI was higher than 28 kg/m2. The incidence of OASIS was not higher in women with episiotomy or vacuum extraction, but it was higher in oxytocin augmentation (P textless 0.031). Conclusion The findings can assist in identification of pregnant women with a higher risk of OASIS who require special attention at delivery to prevent it. In high risk women EUS is indicated to identify and treat possible OASIS as early as possible in order to prevent anal incontinence. |
Zdravevski, E; Lameski, P; Trajkovik, V; Kulakov, A; Chorbev, I; Goleva, R; Pombo, N; Garcia, N Improving Activity Recognition Accuracy in Ambient-Assisted Living Systems by Automated Feature Engineering Journal Article In: IEEE Access, vol. 5, pp. 5262-5280, 2017, ISSN: 2169-3536. @article{zdravevski_Access2017, Ambient-assisted living (AAL) is promising to become a supplement of the current care models, providing enhanced living experience to people within context-aware homes and smart environments. Activity recognition based on sensory data in AAL systems is an important task because 1) it can be used for estimation of levels of physical activity, 2) it can lead to detecting changes of daily patterns that may indicate an emerging medical condition, or 3) it can be used for detection of accidents and emergencies. To be accepted, AAL systems must be affordable while providing reliable performance. These two factors hugely depend on optimizing the number of utilized sensors and extracting robust features from them. This paper proposes a generic feature engineering method for selecting robust features from a variety of sensors, which can be used for generating reliable classification models. From the originally recorded time series and some newly generated time series [i.e., magnitudes, first derivatives, delta series, and fast Fourier transformation (FFT)-based series], a variety of time and frequency domain features are extracted. Then, using two-phase feature selection, the number of generated features is greatly reduced. Finally, different classification models are trained and evaluated on an independent test set. The proposed method was evaluated on five publicly available data sets, and on all of them, it yielded better accuracy than when using hand-tailored features. The benefits of the proposed systematic feature engineering method are quickly discovering good feature sets for any given task than manually finding ones suitable for a particular task, selecting a small feature set that outperforms manually determined features in both execution time and accuracy, and identification of relevant sensor types and body locations automatically. Ultimately, the proposed method could reduce the cost of AAL systems by facilitating execution of algorithms on devices with limited- resources and by using as few sensors as possible. |
Lameski, P; Zdravevski, E; Koceski, S; Kulakov, A; Trajkovik, V Suppression of Intensive Care Unit False Alarms based on the Arterial Blood Pressure Signal Journal Article In: IEEE Access, vol. PP, no. 99, pp. 1-7, 2017, ISSN: 2169-3536, (in press). @article{lameski_Access2017, Patient monitoring in Intensive Care Units (ICU) requires collection and processing of high volumes of data. High sensitivity of sensors leads to significant number of false alarms which cause alarm fatigue. Reduction of false alarms can lead to better reaction time of medical personnel. This study aims to develop a method for false alarm suppression and evaluate it on a publicly available dataset with manually annotated alarms. First, an automated feature engineering was performed using the signal for Arterial Blood Pressure (ABP) and a processed signal that contained the times of each heartbeat from the ABP signal. Next, Support Vector Machines, Random Forest and Extreme Random Trees classifiers were trained to create classification models. The best suppression performance was achieved for the Extreme tachycardia alarm, for which we 90.3% of the false alarms were suppressed, while only 0.54% of the true alarms were incorrectly suppressed. This study demonstrates that alarm suppression can be achieved with high accuracy using on automated feature engineering coupled with machine learning algorithms. The proposed approach can be utilized as aid to medical personnel and experts, allowing them to be more productive and to respond to alarms in a more timely manner |
Zdravevski, Eftim; Stojkoska, Biljana Risteska; Standl, Marie; Schulz, Holger Automatic machine-learning based identification of jogging periods from accelerometer measurements of adolescents under field conditions Journal Article In: PLOS ONE, vol. 12, no. 9, pp. 1-28, 2017. @article{zdravevski2017jogging, Background Assessment of health benefits associated with physical activity depend on the activity duration, intensity and frequency, therefore their correct identification is very valuable and important in epidemiological and clinical studies. The aims of this study are: to develop an algorithm for automatic identification of intended jogging periods; and to assess whether the identification performance is improved when using two accelerometers at the hip and ankle, compared to when using only one at either position. Methods The study used diarized jogging periods and the corresponding accelerometer data from thirty-nine, 15-year-old adolescents, collected under field conditions, as part of the GINIplus study. The data was obtained from two accelerometers placed at the hip and ankle. Automated feature engineering technique was performed to extract features from the raw accelerometer readings and to select a subset of the most significant features. Four machine learning algorithms were used for classification: Logistic regression, Support Vector Machines, Random Forest and Extremely Randomized Trees. Classification was performed using only data from the hip accelerometer, using only data from ankle accelerometer and using data from both accelerometers. Results The reported jogging periods were verified by visual inspection and used as golden standard. After the feature selection and tuning of the classification algorithms, all options provided a classification accuracy of at least 0.99, independent of the applied segmentation strategy with sliding windows of either 60s or 180s. The best matching ratio, i.e. the length of correctly identified jogging periods related to the total time including the missed ones, was up to 0.875. It could be additionally improved up to 0.967 by application of post-classification rules, which considered the duration of breaks and jogging periods. There was no obvious benefit of using two accelerometers, rather almost the same performance could be achieved from either accelerometer position. Conclusions Machine learning techniques can be used for automatic activity recognition, as they provide very accurate activity recognition, significantly more accurate than when keeping a diary. Identification of jogging periods in adolescents can be performed using only one accelerometer. Performance-wise there is no significant benefit from using accelerometers on both locations. |
Book Chapters
2020 |
Pires, Ivan Miguel; Marques, Gonçalo; Garcia, Nuno M; Pombo, Nuno; Flórez-Revuelta, Francisco; Zdravevski, Eftim; Spinsante, Susanna In: Singh, Pradeep Kumar; Bhargava, Bharat K; Paprzycki, Marcin; Kaushal, Narottam Chand; Hong, Wei-Chiang (Ed.): Handbook of Wireless Sensor Networks: Issues and Challenges in Current Scenario's, pp. 685–713, Springer International Publishing, Cham, 2020, ISBN: 978-3-030-40305-8. @inbook{Pires2020, Smart environments and mobile devices are two technologies that when combined may allow the recognition of Activities of Daily Living (ADL) and its environments. This paper focuses on the literature review of the existing machine learning methods for the recognition of ADL and its environments, by means of comparison jointly with a proposal of a novel taxonomy in this context. The sensors used for this purpose depends on the nature of the system and the ADL to recognize. The available in the mobile devices are mainly motion, magnetic and location sensors, but the sensors available in the smart environments may have different types. Data acquired from several sensors can be used for the identification of ADL, where the motion, magnetic and location sensors handle the recognition of activities with movement, and the acoustic sensors handle the recognition of activities related with the environment. |
Pereira, Gonçalo F Valentim; Pires, Ivan Miguel; Marques, Gonçalo; Garcia, Nuno M; Zdravevski, Eftim; Lameski, Petre; Flórez-Revuelta, Francisco; Spinsante, Susanna Mobile Applications Dedicated for Cardiac Patients: Research of Available Resources Book Chapter In: Balas, Valentina E; Solanki, Vijender Kumar; Kumar, Raghvendra (Ed.): Internet of Things and Big Data Applications: Recent Advances and Challenges, pp. 107–119, Springer International Publishing, Cham, 2020, ISBN: 978-3-030-39119-5. @inbook{ValentimPereira2020, In recent years cardiac problems and using mobile devices for aiding people with these problems have received significant attention from the scientific communities to develop solutions to improve the quality of life. The proliferation of mobile computing technologies has revolutionized the medical practices in both patient and clinical staff sides. In particular, the development of mobile health applications continues to increase; mainly, the cardiology field is the most addressed. This paper focuses on the review of the mobile applications available in the Google Play Store that are dedicated to cardiac patients. The number of cardiac patients is increasing, but there are no mobile applications that aid cardiac patients by providing monitoring of different parameters, including the calorie intake and the calories burned. However, the mobile applications that can be adapted to this type of people were analyzed. We found six notable mobile applications. Their features can be grouped in diet, anthropometric parameters, and physical activity. |
2019 |
Zdravevski, Eftim; Lameski, Petre; Trajkovik, Vladimir; Chorbev, Ivan; Goleva, Rossitza; Pombo, Nuno; Garcia, Nuno M Automation in Systematic, Scoping and Rapid Reviews by an NLP Toolkit: A Case Study in Enhanced Living Environments Book Chapter In: Ganchev, Ivan; Garcia, Nuno M; Dobre, Ciprian; Mavromoustakis, Constandinos X; Goleva, Rossitza (Ed.): Enhanced Living Environments: Algorithms, Architectures, Platforms, and Systems, pp. 1–18, Springer International Publishing, Cham, 2019, ISBN: 978-3-030-10752-9. @inbook{Zdravevski2019c, With the increasing number of scientific publications, the analysis of the trends and the state-of-the-art in a certain scientific field is becoming very time-consuming and tedious task. In response to urgent needs of information, for which the existing systematic review model does not well, several other review types have emerged, namely the rapid review and scoping reviews. In this paper, we propose an NLP powered tool that automates most of the review process by automatic analysis of articles indexed in the IEEE Xplore, PubMed, and Springer digital libraries. We demonstrate the applicability of the toolkit by analyzing articles related to Enhanced Living Environments and Ambient Assisted Living, in accordance with the PRISMA surveying methodology. The relevant articles were processed by the NLP toolkit to identify articles that contain up to 20 properties clustered into 4 logical groups. The analysis showed increasing attention from the scientific communities towards Enhanced and Assisted living environments over the last 10 years and showed several trends in the specific research topics that fall into this scope. The case study demonstrates that the NLP toolkit can ease and speed up the review process and show valuable insights from the surveyed articles even without manually reading of most of the articles. Moreover, it pinpoints the most relevant articles which contain more properties and therefore, significantly reduces the manual work, while also generating informative tables, charts and graphs. |
2017 |
Lameski, Petre; Zdravevski, Eftim; Trajkovik, Vladimir; Kulakov, Andrea Weed Detection Dataset with RGB Images Taken Under Variable Light Conditions Book Chapter In: Trajanov, Dimitar; Bakeva, Verica (Ed.): ICT Innovations 2017: Data-Driven Innovation. 9th International Conference, ICT Innovations 2017, Skopje, Macedonia, September 18-23, 2017, Proceedings, pp. 112–119, Springer International Publishing, Cham, 2017, ISBN: 978-3-319-67597-8. @inbook{Lameski2017b, Weed detection from images has received a great interest from scientific communities in recent years. However, there are only a few available datasets that can be used for weed detection from unmanned and other ground vehicles and systems. In this paper we present a new dataset (i.e. Carrot-Weed) for weed detection taken under variable light conditions. The dataset contains RGB images from young carrot seedlings taken during the period of February in the area around Negotino, Republic of Macedonia. We performed initial analysis of the dataset and report the initial results, obtained using convolutional neural network architectures. |
2010 |
Zdravevski, Eftim; Kulakov, Andrea System for Prediction of the Winner in a Sports Game Book Chapter In: pp. 55–63, Springer, 2010. @inbook{zdravevski2010system, |
International Conferences
2019 |
Zdravevski, Ace Dimitrievski Petre Lameski Eftim Cluster-size optimization within a cloud-based ETL framework for Big Data Inproceedings In: Proceedings of the 2019 IEEE Big Data Conference - IEEE BigDATA 2019, IEEE IEEE, Los Angeles, USA, 2019. @inproceedings{zdravevski2019cluster, |
Apanowicz, Krzysztof Stencel Eftim Zdravevski Cas; Slezak, Dominik Scalable Cloud-based ETL for Self-serving Analytics Inproceedings In: Perner, Petra (Ed.): Proceedings of the 19th Industrial Conference on Data Mining ICDM 2019, pp. 317–331, IBAI Publishing, New York, USA, 2019. @inproceedings{zdravevski2019icdm, |
Lameski, J; Jovanov, A; Zdravevski, E; Lameski, P; Gievska, S Skin lesion segmentation with deep learning Inproceedings In: IEEE EUROCON 2019 -18th International Conference on Smart Technologies, pp. 1-5, 2019, ISSN: null. @inproceedings{Lameski2019skin, |
Lameski, P; Dimitrievski, A; Zdravevski, E; Trajkovik, V; Koceski, S Challenges in data collection in real-world environments for activity recognition Inproceedings In: IEEE EUROCON 2019 -18th International Conference on Smart Technologies, pp. 1-5, 2019, ISSN: null. @inproceedings{Lameski2019challenges, |
Villasana, María Vanessa; Pires, Ivan Miguel; Sá, Juliana; Garcia, Nuno M; Pombo, Nuno; Zdravevski, Eftim; Chorbev, Ivan CoviHealth: Novel Approach of a Mobile Application for Nutrition and Physical Activity Management for Teenagers Inproceedings In: Proceedings of the 5th EAI International Conference on Smart Objects and Technologies for Social Good, pp. 261–266, ACM, Valencia, Spain, 2019, ISBN: 978-1-4503-6261-0. @inproceedings{Villasana2019, |
Dimitrievski, Ace; Zdravevski, Eftim; Lameski, Petre; Trajkovik, Vladimir Addressing Privacy and Security in Connected Health with Fog Computing Inproceedings In: Proceedings of the 5th EAI International Conference on Smart Objects and Technologies for Social Good, pp. 255–260, ACM, Valencia, Spain, 2019, ISBN: 978-1-4503-6261-0. @inproceedings{Dimitrievski2019addressing, |
2018 |
Lameski, Petre; Zdravevski, Eftim; Kalajdziski, Slobodan; Trajkova, Vesna; Hadzieva, Elena Computer-Aided Detection of Melanoma, A Case Study Inproceedings In: Proceedings of ETAI, IKT-ACT Struga, Macedonia, 2018. @inproceedings{lameski2018computer, |
Videnovik, M; Zdravevski, E; Lameski, P; Trajkovik, V The BBC Micro:bit in the International Classroom: Learning Experiences and First Impressions Inproceedings In: 2018 17th International Conference on Information Technology Based Higher Education and Training (ITHET), pp. 1-5, 2018, ISSN: null. @inproceedings{videnovik2018microbit, |
Lameski, Petre; Zdravevski, Eftim; Kulakov, Andrea Review of Automated Weed Control Approaches: An Environmental Impact Perspective Inproceedings In: Kalajdziski, Slobodan; Ackovska, Nevena (Ed.): ICT Innovations 2018. Engineering and Life Sciences, pp. 132–147, Springer International Publishing, Cham, 2018, ISBN: 978-3-030-00825-3. @inproceedings{lameski2018weedreview, Agricultural food production is in constant struggle to meet the market demands. Weed control is used to increase the per land unit production from agricultural field. The process of weed removal is usually performed manually and is a time-consuming and labor demanding task. Since mechanical removal is a difficult process, the plantations use herbicides to remove unwanted plants. Herbicides are applied in large quantities, thus often have a degenerative effect on the land. Sometimes, they even endanger the health of the workers who apply them and the end users which consume the harvested product. We review the technologies used for automated weed control and its environmental impact, specifically on the pollution reduction. We also review the herbicides reduction reported in implemented and tested approaches for precision agriculture with emphasis on the weed control environmental impact. Based on the reviewed papers, we conclude that automated weed detection can identify unwanted plants with decent accuracy. Consequently, this can facilitate building autonomous spraying systems that can significantly reduce the quantity of applied herbicides by precisely applying the chemicals only on the plants or mechanically removing unwanted plants. We also review the challenges that need to be overcome, such as precise weed plant type detection, speed of the process and some security considerations that arise from the involvement of information and communication technologies. |
Koteli, Petre Lameski Eftim Zdravevski 12. Vladimir Stojov Nikola Application of machine learning and time-series analysis for air pollution prediction Inproceedings In: Proceedings of the 15th International Conference on Informatics and Information Technologies (CIIT 2018), Faculty of Computer Science and Engineering (FCSE) and Computer Society of Macedonia Mavrovo, Macedonia, 2018. @inproceedings{stojov2018air, |
Zdravevski, Vladimir Trajkovik Andrea Kulakov Petre Lameski Eftim Cloud based Data Acquisition and Annotation Architecture for Weed Control Inproceedings In: Proceedings of the 15th International Conference on Informatics and Information Technologies (CIIT 2018), Faculty of Computer Science and Engineering (FCSE) and Computer Society of Macedonia Mavrovo, Macedonia, 2018. @inproceedings{lameski2018cloudweed, |
Docevski, Marko; Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea Towards Music Generation With Deep Learning Algorithms Inproceedings In: Proceedings of the 15th International Conference on Informatics and Information Technologies (CIIT 2018), Faculty of Computer Science and Engineering (FCSE) and Computer Society of Macedonia Mavrovo, Macedonia, 2018. @inproceedings{docevski2018towards, |
2017 |
Alla, Arijan; Zdravevski, Eftim; Trajkovik, Vladimir Framework for aiding surveys by natural language processing Inproceedings In: 2017 ICT Innovations web proceedings, 2017. @inproceedings{alla2017framework, |
Zdravevska, Aleksandra; Dimitrievski, Ace; Lameski, Petre; Zdravevski, Eftim; Trajkovik, Vladimir Cloud-based Privacy Preserving Recognition of Complex Activities for Ambient Assisted Living in Smart Homes Inproceedings In: Proceedings of the 17th IEEE International Conference on Smart Technologies IEEE EUROCON 2017, IEEE, Ohrid, Macedonia, 2017, (In press). @inproceedings{zdravevska2017, |
Lameski, Petre; Zdravevski, Eftim; Kulakov, Andrea; Trajkovik, Vladimir Cloud Based Architecture for Automated Weed Control Inproceedings In: Proceedings of the 17th IEEE International Conference on Smart Technologies IEEE EUROCON 2017, IEEE, Ohrid, Macedonia, 2017, (In press). @inproceedings{lameski2017cloud, |
2016 |
Zdravevski, E; Lameski, P; Kulakov, A Automatic Feature Engineering for Prediction of Dangerous Seismic Activities in Coal Mines Inproceedings In: Maciaszek, Paprzycki Ganzha M M L (Ed.): Computer Science and Information Systems (FedCSIS), 2016 Federated Conference on, pp. 251–254, IEEE, 2016. @inproceedings{zdravevski_fedcsis_DMC_2016, |
Dimitrievski, A; Zdravevski, E; Lameski, P; Trajkovik, V A survey of Ambient Assisted Living systems: Challenges and opportunities Inproceedings In: 2016 IEEE 12th International Conference on Intelligent Computer Communication and Processing (ICCP), pp. 49-53, 2016. @inproceedings{dimitrievski2016survey, As the research in Ambient Assisted Living (AAL) matures, we expect that data generated from AAL IoT devices will benefit from analysis by well established machine learning techniques. There is also potential that new research in ML and Artificial Intelligence (AI) can be used on data generated from the sensors used in AAL. In this paper we present a survey of the research in the related topics, identify its shortcomings and propose future work that will integrate these fields by collecting ambient sensor data and process the data by ML framework which can detect and classify activities. |
Dimitrievski, Ace; Zdravevski, Eftim; Lameski, Petre; Trajkovik, Vladimir Towards application of non-invasive environmental sensors for risks and activity detection Inproceedings In: 2016 IEEE 12th International Conference on Intelligent Computer Communication and Processing (ICCP), pp. 27-33, 2016. @inproceedings{dimitrievski2016application, |
Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea; Trajkovik, Vladimir Performance Comparison of Random Forests and Extremely Randomized Trees Inproceedings In: Proceedings of the 13th Conference for Informatics and Information Technology (CIIT 2016), Faculty of Computer Science and Engineering (FCSE) and Computer Society of Macedonia Bitola, Macedonia, 2016. @inproceedings{zdravevski2016performance, |
Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea Row Key Designs of NoSQL Database Tables and Their Impact on Write Performance Inproceedings In: Proceedings - 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016, pp. 10–17, IEEE, 2016, ISBN: 9781467387750. @inproceedings{zdravevski2016nosql, In several NoSQL database systems, among which is HBase, only one index is available for the tables, which is also the row key and the clustered index. Using other indexes does not come out of the box. As a result, the row key design is the most important thing when designing tables, because an inappropriate design can lead to detrimental consequences on performances and costs. Particular row key designs are suitable for different problems, and in this paper we analyze the performance, characteristics and applicability of each of them. In particular we investigate the effect of using various techniques for modeling row keys: sequences, salting, padding, hashing, and modulo operations. We propose four different designs based on these techniques and we analyze their performance on different HBase clusters when loading HDFS files with various sizes. The experiments show that particular designs consistently outperform others on differently sized clusters in both execution time and even load distribution across nodes. |
Mingov, Riste; Zdravevski, Eftim; Lameski, Petre Application of Russian Language Phonemics to Generate Macedonian Speech Recognition Model Using Sphinx Inproceedings In: ICT Innovations 2016, Web Proceedings, IKT-ACT 2016. @inproceedings{mingov2016language, |
2015 |
Zdravevski, E; Lameski, P; Mingov, R; Kulakov, A; Gjorgjevikj, D Robust histogram-based feature engineering of time series data Inproceedings In: Maciaszek, Paprzycki Ganzha M M L (Ed.): Computer Science and Information Systems (FedCSIS), 2015 Federated Conference on, pp. 381–388, IEEE, 2015. @inproceedings{zdravevski_fedcsis_DMC_2015, Collecting data at regular time nowadays is ubiquitous. The most widely used type of data that is being collected and analyzed is financial data and sensor readings. Various businesses have realized that financial time series analysis is a powerful analytical tool that can lead to competitive advantages. Likewise, sensor networks generate time series and if they are properly analyzed can give a better understanding of the processes that are being monitored. In this paper we propose a novel generic histogram-based method for feature engineering of time series data. The preprocessing phase consists of several steps: deseansonalyzing the time series data, modeling the speed of change with first derivatives, and finally calculating histograms. By doing all of those steps the goal is three-fold: achieve invariance to different factors, good modeling of the data and preform significant feature reduction. This method was applied to the AAIA Data Mining Competition 2015, which was concerned with recognition of activities carried out by firefighters by analyzing body sensor network readings. By doing that we were able to score the third place with predictive accuracy of about 83%, which was about 1% worse than the winning solution. |
Lameski, P; Zdravevski, E; Kulakov, A Unsupervised weed detection in spinach seedling organic farms Inproceedings In: Proceedings of the 24th International Electrotechnical and Computer Science Conference ERK 2015, Portoroz, Slovenia, 2015. @inproceedings{lameski2015weed, |
Zdravevski, E; Lameski, P; Kulakov, A; Jakimovski, B; Filiposka, S; Trajanov, D Feature ranking based on information gain for large classification problems with MapReduce Inproceedings In: Proceedings of the 9th IEEE International Conference on Big Data Science and Engineering, pp. 186–191, IEEE Computer Society Conference Publishing, Helsinki, Finland, 2015. @inproceedings{zdravevski_BigDataSE2015, In classification problems the large number of features can pose a significant challenge from many aspects. This is particularly the case in the context of Big Data. In order to address this issue we propose a distributed and parallel computation of information gain based on MapReduce. The proposed implementation on Hadoop can be used for ranking features of large datasets and furthermore for feature selection. The data-parallelism is achieved by uniformly distributing it using HBase tables with proper row keys. Performance evaluations are made by estimation of the speed-up of multi-node clusters against a one-node cluster. The framework was deployed on a on-premises Hadoop cluster. The results show that by parallelization and distribution of the computations on a cluster significant speedup can be achieved. The main contribution of this paper is that we have demonstrated how the higher level scripting language Pig Latin can be used for writing MapReduce jobs instead of directly writing a separate map and reduce function. Additionally, we have proposed the use of manually pre-splitted HBase tables instead of HDFS files for data fragmentation in order to set the degree of parallelism on a higher level. |
Lameski, P; Zdravevski, E; Kulakov, A; Gjorgjevik, D Plant species detection based on leaf contours Inproceedings In: Proceedings of the 12th Conference for Informatics and Information Technology (CIIT 2015), Faculty of Computer Science and Engineering (FCSE) and Computer Society of Macedonia Bitola, Macedonia, 2015. @inproceedings{lameski2015plant, |
Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea; Filiposka, Sonja; Trajanov, Dimitar Simplifying parallel implementation of algorithms on Hadoop with Pig Latin Inproceedings In: Proceedings of the 12th Conference for Informatics and Information Technology (CIIT 2015), Faculty of Computer Science and Engineering (FCSE) and Computer Society of Macedonia Bitola, Macedonia, 2015. @inproceedings{zdravevski2015simplifying, |
Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea; Kalajdziski, Slobodan Transformation of nominal features into numeric in supervised multi-class problems based on the weight of evidence parameter Inproceedings In: Maciaszek, Paprzycki Ganzha M M L (Ed.): Proceedings of the 2015 Federated Conference on Computer Science and Information Systems, pp. 169–179, IEEE, 2015. @inproceedings{zdravevski_fedcsis_woe_2015, |
Lameski, Petre; Zdravevski, Eftim; Mingov, Riste; Kulakov, Andrea SVM Parameter Tuning with Grid Search and Its Impact on Reduction of Model Over-fitting Inproceedings In: Yao, Yiyu; Hu, Qinghua; Yu, Hong; Grzymala-Busse, Jerzy W (Ed.): Rough Sets, Fuzzy Sets, Data Mining, and Granular Computing: 15th International Conference, RSFDGrC 2015, Tianjin, China, November 20-23, 2015, Proceedings, pp. 464–474, Springer International Publishing, Cham, Switzerland, 2015, ISBN: 978-3-319-25783-9. @inproceedings{lameski2015svm, |
Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea; Filiposka, Sonja; Trajanov, Dimitar; Jakimovski, Boro Parallel computation of information gain using Hadoop and MapReduce Inproceedings In: Maciaszek, Paprzycki Ganzha M M L (Ed.): Proceedings of the 2015 Federated Conference on Computer Science and Information Systems, pp. 181–192, IEEE, 2015. @inproceedings{zdravevski_fedcsis_InfoGain_2015, |
2014 |
Lameski, Petre; Kulakov, Darko; Zdravevski, Eftim; Kulakov, Andrea Tumor detection in manually selected regions of MRI images Inproceedings In: ICT Innovations 2014, Web Proceedings ISSN 1857-7288, pp. 183-190, IKT-ACT 2014. @inproceedings{lameski2014, |
Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea; Gjorgjevikj, Dejan Feature selection and allocation to diverse subsets for multi-label learning problems with large datasets Inproceedings In: Maciaszek, Paprzycki Ganzha M M L (Ed.): Proceedings of the 2014 Federated Conference on Computer Science and Information Systems, pp. 387–394, IEEE, 2014. @inproceedings{zdravevski_fedcsis2014, Feature selection is important phase in machine learning and in the case of multi-label classification, it can be considerably challenging. In like manner, finding the best subset of good features is involved and difficult when the dataset has significantly large number of features (more than a thousand). In this paper we address the problem of feature selection for multi-label classification with large number of features. The proposed method is a hybrid of two phases - preliminary feature selection based on the information value and additional correlation-based selection.We show how with the first phase we can do preliminary selection of features from tens of thousands to couple of hundred, and then with the second phase we can make fine-grained feature selection with more sophisticated but computationally intensive methods. Finally, we analyze the ways of allocating the selected features to diverse subsets, which are suitable for training of ensembles of classifiers. |
2013 |
Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea Advanced transformations for nominal and categorical data into numeric data in supervised learning problems Inproceedings In: Mishkovski, Igor; Ristov, Sashko (Ed.): Proceedings of the 10th Conference for Informatics and Information Technology (CIIT 2013), pp. 142–146, Faculty of Computer Science and Engineering (FCSE) and Computer Society of Macedonia Bitola, Macedonia, 2013, ISBN: 978-608-4699-01-9. @inproceedings{zdravevski_ciit2013, In the last decade machine learning has gained substantial interest in industry and has been applied to almost all areas for which digital data is present. Very often the available data is multivariate and contains both numeric and categorical (i.e. nominal) features. However, many machine-learning algorithms do not natively support categorical features. This is one of the reasons why the data needs to be pre-processed before a machine-learning algorithm can use it. The most common technique for transformation of nominal features into numeric is by generating dummy (binary) variables for all different values of the nominal features. Few of the drawbacks of this technique are that: it does not optimally exploit the predictive potential of the data and it can slow down many algorithms because of the potentially large number of features it can generate. In this paper we present the results of our research based on applying a new technique for data transformation that is based on the weight of evidence (WoE) parameter. We have tested the WoE technique on binary and multiclass classification problems and the results show significant improvements over the technique that generates dummy variables. |
Lameski, Petre; Zdravevski, Eftim; Mingov, Riste; Kulakov, Andrea Comparison of local image descriptors for plant identification from leaf image Inproceedings In: Proceedings of the 2013 ICMER Conference, 2013. @inproceedings{lameski2013comparison, |
Dikovski, Bojan; Lameski, Petre; Zdravevski, Eftim; Kulakov, Andrea Structure from motion obtained from low quality images in indoor environment Inproceedings In: Proceedings of the 10th Conference for Informatics and Information Technology (CIIT 2013), Faculty of Computer Science and Engineering (FCSE) and Computer Society of Macedonia Bitola, Macedonia, 2013. @inproceedings{dikovski2013structure, |
2012 |
Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea Towards a general technique for transformation of nominal features into numeric features in supervised learning Inproceedings In: Proceedings of the 9th Conference for Informatics and Information Technology (CIIT 2012), Faculty of Computer Science and Engineering (FCSE) and Computer Society of Macedonia Bitola, Macedonia, 2012. @inproceedings{zdravevski2012towards, |
Angelovski, Martin; Lameski, Petre; Zdravevski, Eftim; Kulakov, Andrea Application of BCI Technology for Color Prediction Using Brainwaves Inproceedings In: ICT Innovations 2012, Web Proceedings ISSN 1857-7288, pp. 253, IKT-ACT 2012. @inproceedings{angelovski2012application, |
2011 |
Lameski, Petre; Zdravevski, Eftim; Kulakov, Aandrea; Davcev, Danco Architecture for Wireless Sensor and Actor Networks Control and Data Acquisition Inproceedings In: 2011 International Conference on Distributed Computing in Sensor Systems and Workshops (DCOSS), pp. 1-3, 2011, ISSN: 2325-2936. @inproceedings{lameski2011wsn, Wireless Sensor and Actor Networks (WSANs) have received increased attention from the research community in the recent years. This is mainly because as an extension to Wireless Sensor Networks(WSN), they have the ability to actively participate in the environment trough the actors. However, this introduces new challenges as to how to transfer commands between nodes, actors and central station who may be from different manufacturers and use different communication protocols. Another important aspect is the ability of the WSAN to present the data to the interested party or to receive the command from the operator, and do this with in the simplest and most user friendly way as possible. In this paper we propose architecture for interconnection between different layers of WSANs and the central stations that would allow building a simple interface that would ease the operation with WSANs in view of Control and Data Acquisition. |
Zdravevski, Eftim; Kulakov, Andrea; Kalajdziski, Slobodan; Davcev, Danco Probabilistic Predictions of Ensemble of Classifiers Combined With Dynamically Weighted Majority Vote Inproceedings In: Proceedings of Artificial Intelligence and Applications 2011 (AIA 2011), pp. 236-244, IASTED 2011. @inproceedings{zdravevski2011probabilistic, |
Zdravevski, Eftim; Lameski, Petre; Kulakov, Andrea Weight of evidence as a tool for attribute transformation in the preprocessing stage of supervised learning algorithms Inproceedings In: Neural Networks (IJCNN), The 2011 International Joint Conference on, pp. 181–188, IEEE IEEE, 2011. @inproceedings{zdravevski2011weightb, |
Contact Us form
We are pleased to receive your request, analyze it and then come back to you with a proposal.
