Clusters Of Items Research Articles

Improved Hybrid Collaborative Fitering Algorithm Based on Spark Platform

An improved Hybrid Collaborative Filtering algorithm (H-CF) is proposed, addressing the issues of data sparsity, low recommendation accuracy, and poor scalability present in traditional collaborative filtering algorithms. The core of H-CF is a linear weighted hybrid algorithm based on the Latent Factor Model (LFM) and the Improved Item Clustering and Similarity Calculation Collaborative Filtering Algorithm (ITCSCF). To begin with, the items are clustered based on their attribute dimension, which accelerates the computation of the nearest neighbor set. Subsequently, H-CF enhances the formula for scoring similarity by penalizing popular items and optimizing unpopular items. This improvement enhances the rationality of scoring similarity and reduces the impact of data sparseness. Furthermore, a weighting function is employed to combine the various improved algorithms. The balance factor of the weighting function is dynamically adjusted to attain the optimal recommendation list. To address the real-time and scalability concerns, the algorithm leverages the Spark big data distributed cluster computing framework. Experiments were conducted using the public dataset MovieLens, where the improved algorithm's performance was compared against the algorithm before enhancement and the algorithm running on a single machine. The experimental results demonstrate that the improved algorithm outperforms in terms of data sparsity, recommendation personalization, accuracy, recall, and efficiency.

Interview with Chris Taylor, curator of Shifting Borders: A Journey to the Centre of Our World(s)1

Shifting Borders exhibits ‘traces’ of human mobility across millennia: of exploratory journeys, forced migrations, exilic arrivals at foreign lands and artistic schematizations of them based on memory and experience. The traces assume the form of maps, passports, photographs of people and locations, as well as actual diaries of movement produced by those who move or by institutions who regulate their movements. Crafted as a form of pilgrimage to these stories, the exhibition transcends binary understandings of hospitality as an inviolable norm and/or a secular pact conforming with the commercial rules of catering for strangers. Instead, the presentation of items in clusters produces variations of story-telling as a tribute to the presence of human otherness. Featuring styles of inscription and creative staging of particular mobility events, Shifting Borders showcases forms of movement vis-à-vis (‘psychic centres’): homes and homelands, memories of uprooting but also the excitement of travel and exploration in and of unknown territories. The exhibition’s simulation of such movements transforms artistic pilgrimage to a method of awakening conscience in regard to offering hospitality to others. The exhibition is hosted in the Parkinson Building, one of the heritage buildings of the University of Leeds, as part of the Brotherton Collections, curated by the University’s Libraries. Its rhizomatic story-telling of imaginaries of movement and homemaking reflects the overlapping biographies of its physical location: Leeds as a multicultural city with diverse migrant, exiled and diasporic communities, but also one of the foremost creative cities in the United Kingdom, and the University of Leeds as a pedagogical hub that hosts very diverse student populations from around the world.

Course of Subtypes of Late-Life Depression Identified by Bipartite Network Analysis During Psychosocial Interventions

Approximately half of older adults with depression remain symptomatic at treatment end. Identifying discrete clinical profiles associated with treatment outcomes may guide development of personalized psychosocial interventions. To identify clinical subtypes of late-life depression and examine their depression trajectory during psychosocial interventions in older adults with depression. This prognostic study included older adults aged 60 years or older who had major depression and participated in 1 of 4 randomized clinical trials of psychosocial interventions for late-life depression. Participants were recruited from the community and outpatient services of Weill Cornell Medicine and the University of California, San Francisco, between March 2002 and April 2013. Data were analyzed from February 2019 to February 2023. Participants received 8 to 14 sessions of (1) personalized intervention for patients with major depression and chronic obstructive pulmonary disease, (2) problem-solving therapy, (3) supportive therapy, or (4) active comparison conditions (treatment as usual or case management). The main outcome was the trajectory of depression severity, assessed using the Hamilton Depression Rating Scale (HAM-D). A data-driven, unsupervised, hierarchical clustering of HAM-D items at baseline was conducted to detect clusters of depressive symptoms. A bipartite network analysis was used to identify clinical subtypes at baseline, accounting for both between- and within-patient variability across domains of psychopathology, social support, cognitive impairment, and disability. The trajectories of depression severity in the identified subtypes were compared using mixed-effects models, and time to remission (HAM-D score ≤10) was compared using survival analysis. The bipartite network analysis, which included 535 older adults with major depression (mean [SD] age, 72.7 [8.7] years; 70.7% female), identified 3 clinical subtypes: (1) individuals with severe depression and a large social network; (2) older, educated individuals experiencing strong social support and social interactions; and (3) individuals with disability. There was a significant difference in depression trajectories (F2,2976.9 = 9.4; P < .001) and remission rate (log-rank χ22 = 18.2; P < .001) across clinical subtypes. Subtype 2 had the steepest depression trajectory and highest likelihood of remission regardless of the intervention, while subtype 1 had the poorest depression trajectory. In this prognostic study, bipartite network clustering identified 3 subtypes of late-life depression. Knowledge of patients' clinical characteristics may inform treatment selection. Identification of discrete subtypes of late-life depression may stimulate the development of novel, streamlined interventions targeting the clinical vulnerabilities of each subtype.

The network structure of dysfunctional metacognitions, CAS strategies, and symptoms

In the metacognitive model of psychological disorders, metacognitive strategies and corresponding underlying metacognitive beliefs intensify and maintain emotional distress symptoms. In the current study, our three objectives were to evaluate and replicate the network structure of dysfunctional metacognitions as assessed with the MCQ-30, to examine its stability when adding relevant covariates in the form of metacognitive strategies (worry and rumination) and symptoms (anxiety and depression), and to evaluate how different sets of dysfunctional metacognitions are more or less strongly linked differently to metacognitive strategies and symptoms. A cross-sectional university sample with a mean age of 26 years (N = 440; Males = 156, Females = 283) completed the Metacognitions Questionnaire–30, Penn State Worry Questionnaire, Ruminative Response Scale, and Hopkins Symptom Checklist. Data were analysed using psychological network analysis in R-studio statistical software. The network structure of dysfunctional metacognitions replicated well with item clusters that correspond to clinically meaningful substructures in the metacognitive model. Negative metacognitive beliefs and beliefs about uncontrollability might have more functional significance in the mutual connections between dysfunctional meta-domains as well as the connections with metacognitive strategies and symptoms. For worry and anxiety, negative beliefs about uncontrollability and corresponding danger of worry were more prominently connected in the network structure. For rumination, cognitive self-consciousness was more prominent, whereas for depression, need for control was more prominently connected. Support was found for mutual interdependence between different sets of dysfunctional metacognitive beliefs, that metacognitive beliefs are linked to but separate from metacognitive strategies, and that these may function together in affecting emotional distress symptoms

The Impact of Missing Data on the Performances of DIF Detection Methods

This study analyzed the impact of missing data techniques on performances of two differential item functioning (DIF) detection methods (Mantel Haenszel and Multiple Indicator and Multiple Causes) under missing completely at random missing data mechanism. Percentage of missing data was set at 5% and 15%. Zero imputation, listwise deletion and fractional hot-deck imputation were used to handle missing data. The data set of the study consisted of 17 items in the S12 item cluster of Programme for International Student Assessment (PISA) 2015 science test. Results showed that fractional hot-deck imputation produced the best results in identifying DIF items in all conditions and it had also the closest DIF values to the values obtained from complete data set. It was also found that multiple indicator and multiple causes method was more adversely affected than Mantel Haenszel by the presence of missing data.

Taking Inventory of the Creative Behavior Inventory: An Item Response Theory Analysis of the CBI

ABSTRACT The original 90-item Creative Behavior Inventory (CBI) was a landmark self-report scale in creativity research, and the 28-item brief form developed nearly 20 years ago continues to be a popular measure of everyday creativity. Relatively little is known, however, about the psychometric properties of this widely used scale. In the current research, we conduct a detailed psychometric investigation into the 28-item CBI by applying methods from item response theory using a sample of 2,082 adults. Our investigation revealed several strengths of the current scale: excellent reliability, suitable dimensionality, appropriate item difficulty, and reasonably good item discrimination. Several areas for improvement were highlighted as well: (1) the four-point response scale should have fewer options; (2) a handful of items showed gender-based differential item functioning, indicating some gender bias; and (3) local dependence statistics revealed clusters of items that are redundant and could be trimmed. These analyses support the continued use of the CBI for assessing engagement in everyday creative behaviors but suggest that the CBI could benefit from thoughtful revision.

Psychometric assessment of the Polish Translation of the Transgender Positive Identity Measure (T-PIM).

Research conducted in transgender and gender diverse individuals focused mainly on the challenges and health disparities affecting this population. One reason for this situation is the lack of questionnaires capturing positive aspects and experiences related to being a transgender person. The Transgender Positive Identity Measure (T-PIM) is one of a very few measures designed to explore such experiences. The aim of the study was to investigate the structure, reliability and validity of the Polish translation of the T-PIM questionnaire. A group of 89 transgender and gender diverse participants completed an online survey including the T-PIM questionnaire, resilience measurement scale (SPP-25) and depression (CESD-R) measures. Hierarchical item clustering method (ICLUST), Horn's parallel analysis and Velicer's minimum average partial test (MAP) were employed to investigate the structure of the questionnaire. The analyses showed that the Polish translation of the T-PIM questionnaire was characterized by a 5-factor structure consistent with the original publication (Authenticity, Intimacy, Community, Social Justice, Insights). Cronbach's alpha and Guttman's lambda-6 reliability coefficients reached satisfactory levels for all five factors and for the whole questionnaire. The Polish translation of the T-PIM questionnaire is characterized by satisfying psychometric properties and can be used in studies on transgender and gender diverse communities.

Assessment of patient life engagement in major depressive disorder using items from the Inventory of Depressive Symptomatology Self-Report (IDS-SR)

BackgroundPatient-reported outcomes can measure domains that are personally meaningful, such as life engagement, which reflects motivation, pleasure, and well-being. This study explored whether certain items from the Inventory of Depressive Symptomatology Self-Report (IDS-SR) can capture patient life engagement in major depressive disorder (MDD). MethodsIDS-SR life engagement items were identified by a) a panel of expert psychiatrists (n = 4), b) patient interviews (n = 20), and c) a principal component analysis (PCA) to explore clustering of items. Psychometric analyses were performed on potential subscales, and a minimal clinically important difference (MCID) was estimated by anchor- and distribution-based methods. IDS-SR data were obtained from three randomized controlled trials of adjunctive brexpiprazole in MDD. ResultsExpert psychiatrists selected 10 items by consensus from the IDS-SR that might capture patient life engagement (Cronbach's alpha, 0.82; item–total correlations, 0.36–0.58). Patient interviews identified 13 items as moderately to very relevant to life engagement (Cronbach's alpha, 0.85; item–total correlations, 0.35–0.61). The PCA revealed a cluster that included all 10 items selected by psychiatrists and 11 items identified by patients. Expert psychiatrists intentionally distinguished life engagement and core depressive symptoms, although patient insights and the PCA indicated that these aspects of MDD are strongly linked. The 10-item IDS-SR life engagement subscale had an MCID of 3–5 points. ConclusionsDifferent approaches consistently identified a subset of 10 IDS-SR items that can measure life engagement in MDD, which may be suitable to group into an IDS-SR life engagement subscale.

GaVe: A Webcam-Based Gaze Vending Interface Using One-Point Calibration.

Gaze input, i.e., information input via eye of users, represents a promising method for contact- free interaction in human-machine systems. In this paper, we present the GazeVending interface (GaVe), which lets users control actions on a display with their eyes. The interface works on a regular webcam, available on most of today's laptops, and only requires a short one-point calibration before use. GaVe is designed in a hierarchical structure, presenting broad item cluster to users first and subsequently guiding them through another selection round, which allows the presentation of a large number of items. Cluster/item selection in GaVe is based on the dwell time, i.e., the time duration that users look at a given Cluster/ item. A user study (N=22) was conducted to test optimal dwell time thresholds and comfortable human-to-display distances. Users' perception of the system, as well as error rates and task completion time were registered. We found that all participants were able to quickly understand and know how to interact with the interface, and showed good performance, selecting a target item within a group of 12 items in 6.76 seconds on average. We provide design guidelines for GaVe and discuss the potentials of the system.

Challenges in pastoral care practice in euthanasia: A concept mapping study among Dutch protestant pastors

This study explores the various difficulties that protestant pastors experience in the context of care for parishioners with a wish for euthanasia. In a reproducible and non-subjective way, using the concept mapping method, we cluster and rank-order 600 elements that pastors had mentioned as “difficult in caring for parishioners with a death wish” in an earlier survey study. The three clusters of items ranked as most difficult are connected with inappropriate care, doubts whether euthanasia in specific circumstances is justified, and disagreements between pastor and parishioner, including the emotional burden arising from those disagreements.

A Conceptual and Computational Framework for Aspect-Based Collaborative Filtering Recommender Systems

Many datasets in E-commerce have rich information about items and users who purchase or rate them. This information can enable advanced machine learning algorithms to extract and assign user sentiments to various aspects of the items thus leading to more sophisticated and justifiable recommendations. However, most Collaborative Filtering (CF) techniques rely mainly on the overall preferences of users toward items only. And there is lack of conceptual and computational framework that enables an understandable aspect-based AI approach to recommending items to users. In this paper, we propose concepts and computational tools that can sharpen the logic of recommendations and that rely on users’ sentiments along various aspects of items. These concepts include: The sentiment of a user towards a specific aspect of a specific item, the emphasis that a given user places on a specific aspect in general, the popularity and controversy of an aspect among groups of users, clusters of users emphasizing a given aspect, clusters of items that are popular among a group of users and so forth. The framework introduced in this study is developed in terms of user emphasis, aspect popularity, aspect controversy, and users and items similarity. Towards this end, we introduce the Aspect-Based Collaborative Filtering Toolbox (ABCFT), where the tools are all developed based on the three-index sentiment tensor with the indices being the user, item, and aspect. The toolbox computes solutions to the questions alluded to above. We illustrate the methodology using a hotel review dataset having around 6000 users, 400 hotels and 6 aspects.

Uncovering psychiatric phenotypes using unsupervised machine learning: A data-driven symptoms approach.

Current categorical classification systems of psychiatric diagnoses lead to heterogeneity of symptoms within disorders and common co-occurrence of disorders. We investigated the heterogeneous and overlapping nature of symptom endorsement in a population-based sample across three of the most common categories of psychiatric disorders: depressive disorders, anxiety disorders, and sleep-wake disorders using unsupervised machine learning approaches. We assessed a total of 43 symptoms in a discovery sample of 6,602 participants of the population-based Rotterdam Study between 2009 and 2013, and in a replication sample of 3,005 participants between 2016 and 2020. Symptoms were assessed using the Center for Epidemiologic Studies Depression Scale, the Hospital Anxiety and Depression Scale, and the Pittsburgh Sleep Quality Index. Hierarchical clustering analysis was applied on test items and participants to investigate common patterns of symptoms co-occurrence, and further quantitatively investigated with clustering methods to find groups that may represent similar psychiatric phenotypes. First, clustering analyses of the questionnaire items suggested a three-cluster solution representing clusters of "mixed" symptoms, "depressed affect and nervousness", and "troubled sleep and interpersonal problems". A highly similar clustering solution was independently established in the replication sample. Second, four groups of participants could be separated, and these groups scored differently on the item clusters. We identified three clusters of psychiatric symptoms that most commonly co-occur in a population-based sample. These symptoms clustered stable over samples, but across the topics of depression, anxiety, and poor sleep. We identified four groups of participants that share (sub)clinical symptoms and might benefit from similar prevention or treatment strategies, despite potentially diverging, or lack of, diagnoses.

A Methodology for Clustering Items with Seasonal and Non-seasonal Demand Patterns for Inventory Management

Item clustering has become one of the most important topics in terms of effective inventory management in supply chains. Classification of items in terms of their features, sales or consumption volume and variation is a prerequisite to determine differentiated inventory policies as well as parameters, most common of which is service levels. Volume classification is easily obtained by well-known Pareto approach while coefficient of variance is usu-ally used for variation dimension. Hence, it is not always applicable to classify items under different product families with different demand patterns in terms of variation. In this paper, we propose two algorithms, one based on statistical analysis and the other an unsupervised machine learning algorithm using K-means clustering, both of which differ-entiate seasonal and non-seasonal products where an item’s variation is evaluated with respect to seasonality of the product group it belongs to. We then calculate the efficiency of two proposed approaches by standard deviation within each cluster and absolute difference of percentage of volume and item numbers. We also compare the outputs of two algorithms with the methodology which is based on coefficient of variance and is currently in use at the company which is a leading major domestic appliance manufacturer. The results show that the statistical method we propose generates superior outputs than the other two for both seasonal and non-seasonal demand patterns.

Nongovernmental organizations' leaders' attitude toward social innovation: The validation of a research tool

AbstractThis paper presents the results of a validation study of an inventory to measure the attitude of leaders of nongovernmental organizations (NGOs) toward social innovation related to solving local social problems. The study was conducted using a cross‐validation model and an innovative item cluster analysis method to analyze the data, assessing factor structure, the internal consistency of the test, criterion validity, and stability over time. The results indicate a one‐factor structure for the phenomenon and satisfactory psychometric properties for the inventory. The tool makes possible a study of NGO leaders' attitudes toward social innovation to facilitate the design of educational programs related to solving social problems through social innovation.

Investigation of the Differential Power of Young's Internet Addiction Questionnaire Using the Decision Stump Tree.

Background Internet addiction is one of the serious consequences of recent advances in the use of social media. Early detection of Internet addiction is essential because of its harms and is necessary for timely and effective treatment. Aim The aim of this study was to use data mining and an artificial intelligence algorithm to estimate the differential power of each question in the Young Internet Addiction Test and build a decision stump model to predict which item in the questionnaire can be representative of the whole questionnaire. Methods This is a descriptive study conducted at the University of Tabriz, in which 256 undergraduate students were selected in randomized cluster sampling, and they completed Young's IAT (Internet Addiction Test) questionnaire and some demographic questions. The data were statistically analyzed with SPSS and were divided into two groups, normal and addicted, by using a cut-off point. Also, the data of the subjects was used to model the decision stump tree in WEKA. The clustering item was the normal and addicted specifier. Results The study shows that Cronbach's alpha of the IAT is 0.88, which shows good internal integration of subjects that are used to develop the model in WEKA (the Waikato Environment for Knowledge Analysis). Data analysis showed that by using the second question of this questionnaire as the root of the decision stump tree model, it is possible to distinguish between Internet addicts and healthy users with 82% accuracy using this model. Conclusion The study shows innovative ways in which decision stump trees and data mining can help to improve methods used in Clinical Psychotherapy and Human Science. Regarding this, the study showed that early detection of Internet addiction would be possible by using the 2nd question of the IAT. Also, early detection can result in cost-effectiveness for the whole healthcare system.

Statistical features learning to predict the crop yield in regional areas

&lt;span&gt;The plethora of information presented in the form of benchmark dataset plays a significant role in analyzing and understanding the crop yield in certain regions of regional territory. The information may be presented in the form of attributes makes a prediction of crop yield in various regions of machine learning. The information considered for processing involves data cleaning initially followed by binning to reduce the missing data. The information collected is subjected to clustering of data items based on patterns of similarity, The data items that are similar in nature is fed to the system with similarity measure, which involves understanding the distance of data items from its related data item leading to hyper parameters for analyzing of information while calculating the crop yield. The information may be used to ascertain the patterns of data that exhibit similarity with nearest neighbor represented by another attribute. Thus, the research method has yielded an accuracy of 89.62% of classification for predicting the crop yield in agricultural areas of Karnataka region.&lt;/span&gt;

Psychometric Assessment of Pilot Language and Communication Items on the 2018 and 2019 National Survey of Children's Health.

Until recently, normative data on language and communication development among children in the United States have not been available to inform critical efforts to promote language development and prevent impairments. This study represents the first psychometric assessment of nationally representative data derived from a National Survey of Children's Health (NSCH) pilot measure of language and communication development among children ages 1 to 5 years. We analyzed 14,573 parent responses to language and communication items on the 2018 and 2019 NSCH to evaluate whether the newly added 11 items represent a single latent trait for language and communication development and to determine normative age of success on each item. We applied weighted, one-parameter Item Response Theory to rate and cluster items by difficulty relative to developmental language ability. We examined differential item functioning (DIF) using weighted logistic regression by demographic factors. Together, exploratory factor analysis resulting in a single factor > 1 and explaining 93% of the variance and positive correlations indicated unidimensionality of the measure. Item characteristic curves indicated groupings were overall concordant with proposed milestone ages and representative of an approximate 90% success cut-point by child age. Indicated normative age cut-points for 3 of the items differed slightly from proposed milestone ages. Uniform DIF was not observed and potential nonuniform DIF was observed across 5 items. Results have the potential to enhance understanding of risk and protective factors, inform efforts to promote language and communication development, and guide programmatic efforts on early detection of language delays.

Enhancing Context-Aware Recommendation Using Hesitant Fuzzy Item Clustering by Stacked Autoencoder Based Smoothing Technique

Context-aware recommender systems (CARS) are a key component in businesses, notably in the e-transactions domain, since they assume that reviews, ratings, demographics, and other factors may determine customer preferences. On contrary, while evaluating the sentiment underlying the reviews and the rating score, consumers’ opinion is typically conflicting. As a result, a framework that employs either a review or a rating for top-N recommendation directs to produce unsatisfied recommendations in addition to a meager rating problem and high computation time. To overcome all the problems, a novel sentiment enhanced stacked autoencoder (SSAE) with context-specific hesitant fuzzy item hierarchical clustering (CHFHC) approach is proposed which employs online and offline phases. In the offline-phase, the meager user-item rating matrix is smoothed by learning the users’ concrete preference to a complete matrix by the SSAE approach. They are clustered offline using the CHFHC approach into context-based similar item clusters. In the online-phase, the active user gets context-based recommendations from the most similar cluster that matches the active users’ current context situation. Hence the SSAE_CHFHC approach improves the quality of Top-N recommendation corresponding to the exact contextual situation of the active user with a minimal recommendation computation time. Experiments on the (5-core) Amazon and yelp datasets proved that the intended SSAE_CHFHC framework consistently outperforms state-of-the-art recommendation algorithms on a variety of evaluation measures.

Diabetes Distress, Depressive Symptoms, and Anxiety Symptoms in People With Type 2 Diabetes: A Network Analysis Approach to Understanding Comorbidity

In this study, we aimed to explore interactions between individual items that assess diabetes distress, depressive symptoms, and anxiety symptoms in a cohort of adults with type 2 diabetes using network analysis. Participants (N = 1,796) were from the Montreal Evaluation of Diabetes Treatment (EDIT) study from Quebec, Canada. A network of diabetes distress was estimated using the 17 items of the Diabetes Distress Scale (DDS-17). A second network was estimated using the DDS-17 items, the nine items of the Patient Health Questionnaire (PHQ-9), and the seven items of the Generalized Anxiety Disorder Assessment (GAD-7). Network analysis was used to identify central items, clusters of items, and items that may act as bridges between diabetes distress, depressive symptoms, and anxiety symptoms. Regimen-related and physician-related problems were among the most central (highly connected) and influential (most positive connections) in the diabetes distress network. The depressive symptom of failure was found to be a potential bridge between depression and diabetes distress, being highly connected to diabetes distress items. The anxiety symptoms of worrying too much, uncontrollable worry, and trouble relaxing were identified as bridges linking both anxiety and depressive items and anxiety and diabetes distress items, respectively. Regimen-related and physician-related diabetes-specific problems may be important in contributing to the development and maintenance of diabetes distress. Feelings of failure and worry are potentially strong candidates for explaining comorbidity. These individual diabetes-specific problems and mental health symptoms could hold promise for targeted interventions for people with type 2 diabetes.

Cognitive workload classification: Towards generalization through innovative pipeline interface using HMM

Cognition is an important aspect to realize one’s potential and accelerate progress towards achieving a better quality of life. The cognitive workload is characterized by the occupancy of working memory during task performance. In spite of low spatial resolution, Electroencephalogram (EEG) continues to be a substantial tool in cognitive load research. The current methods use various time and frequency domain features for the assessment of cognitive load levels using EEG. The present work intends to model the temporal dynamics of the EEG signals by proposing a pipeline interface to classify cognitive workload under a continuous engagement/attention environment. The short segments of decomposed data are assumed to be stationary and are modeled using the autoregressive method. Later, fuzzy c-means maps the AR coefficients to low dimensional space. These clustered items are scaled further in the temporal domain by finding the interplay between the successive items through the most likely sequence of hidden states using the hidden Markov model (HMM). The successive segments group, having similar HMM states form variable-length frames. The handcrafted feature set from variable-length frames is classified using a deep recurrent neural network (RNN) structure. Finally, the maximum voting scheme on the predicted RNN output enhances the classification accuracy. The performance of this method is evaluated on a publicly available, and self-collected dataset. The findings reveal that handcrafted features supported by modeling of temporal dynamics outperform state-of-the-art methods. As a result, 97.8% accuracy is observed for multi-class classification with an optimum of four electrodes.