Final Class Research Articles

An Evidential Multi-Target Domain Adaptation Method Based on Weighted Fusion for Cross-Domain Pattern Classification.

For cross-domain pattern classification, the supervised information (i.e., labeled patterns) in the source domain is often employed to help classify the unlabeled target domain patterns. In practice, multiple target domains are usually available. The unlabeled patterns (in different target domains) which have high-confidence predictions, can also provide some pseudo-supervised information for the downstream classification task. The performance in each target domain would be further improved if the pseudo-supervised information in different target domains can be effectively used. To this end, we propose an evidential multi-target domain adaptation (EMDA) method to take full advantage of the useful information in the single-source and multiple target domains. In EMDA, we first align distributions of the source and target domains by reducing maximum mean discrepancy (MMD) and covariance difference across domains. After that, we use the classifier learned by the labeled source domain data to classify query patterns in the target domains. The query patterns with high-confidence predictions are then selected to train a new classifier for yielding an extra piece of soft classification results of query patterns. The two pieces of soft classification results are then combined by evidence theory. In practice, their reliabilities/weights are usually diverse, and an equal treatment of them often yields the unreliable combination result. Thus, we propose to use the distribution discrepancy across domains to estimate their weighting factors, and discount them before fusing. The evidential combination of the two pieces of discounted soft classification results is employed to make the final class decision. The effectiveness of EMDA was verified by comparing with many advanced domain adaptation methods on several cross-domain pattern classification benchmark datasets.

Fostering Engagement Through Ice Breaking in Islamic Education

Background: Innovative pedagogical strategies are crucial for enhancing student engagement in Islamic education. Specific Background: This study examines the implementation of ice-breaking activities to address challenges of fatigue and disengagement during the final class period at an elementary school. Knowledge Gap: There is limited research on the application of ice-breaking techniques within Islamic education contexts. Aims: The research aims to evaluate the effectiveness of ice-breaking activities in fostering a conducive learning environment. Results: Findings indicate that these activities significantly improve students’ attention and comprehension, facilitating effective lesson delivery. Novelty: This study demonstrates how tailored ice-breaking exercises can alleviate classroom fatigue and promote engagement. Implications: The results suggest that integrating ice-breaking activities can serve as a practical strategy for educators to enhance student engagement and knowledge retention, thereby cultivating a dynamic classroom atmosphere. Highlights : Enhancement of Engagement: Ice-breaking activities significantly boost student focus and participation, especially during critical learning periods. Tailored Approaches: Specific ice-breaking techniques can effectively address classroom fatigue and disengagement in Islamic education settings. Practical Teaching Strategy: Integrating these activities offers educators a valuable tool for improving instructional delivery and student comprehension. Keywords: Islamic education, ice-breaking activities, student engagement, classroom strategies, teaching effectiveness.

Health risks from exposure to industrial aerosols of soluble and insoluble nickel compounds

Introduction. Despite the implementation of active preventive measures, employees of nickel enterprises, remain to belong to a group of increased risk of occupational pathology. Aim. comparative analysis of the risks to occupational health in workers exposed to soluble and insoluble nickel compounds in various specialties of the nickel industry. Materials and methods. A retrospective observational study of the health status and development of occupational pathology was carried out among seven hundred sixty three workers of pyrometallurgical and 1395 workers of electrolysis nickel processing plants during 2008–2023. Results. Over the course of 15 years, 112 and 221 occupational diseases were first identified in 57 (7.5%) of 763 pyrometallurgical workers and 100 (7.2%) of 1,395 nickel electrolysis workers. The risk of developing occupational pathology among workers of the two industries did not differ: RR 1.04; CI 0.76-1.43; p=0.797. In 2009–2023 There were no significant trend in the number of workers with primary occupational diseases and their number. Of all the specialists, the greatest risk of occupational pathology was found among finished product cleaners and smelters. The maximum level of occupational morbidity was found among cleaners and smelters (323.62 and 229.69 cases / 10,000 workers), and the minimum – among repairmen, crane operators, roasters (pyrometallurgical production) and electricians of electrolysis production (35–45 cases / 10,000 workers). Limitations. The number of observations in some groups of specialists is insufficient for statistical processing. Conclusion. The risk of developing occupational pathology does not depend on the solubility or insolubility of nickel compounds in industrial aerosols, but is largely determined by the final class of working conditions and the specialty of the worker. When carrying out measures to reduce health risks, it is necessary to take into account not only the class of working conditions, but also their characteristics for specialists at various technological stages of nickel production.

Delving into CLIP latent space for Video Anomaly Recognition

We tackle the complex problem of detecting and recognising anomalies in surveillance videos at the frame level, utilising only video-level supervision. We introduce the novel method AnomalyCLIP, the first to combine Vision and Language Models (VLMs), such as CLIP, with multiple instance learning for joint video anomaly detection and classification. Our approach specifically involves manipulating the latent CLIP feature space to identify the normal event subspace, which in turn allows us to effectively learn text-driven directions for abnormal events. When anomalous frames are projected onto these directions, they exhibit a large feature magnitude if they belong to a particular class. We also leverage a computationally efficient Transformer architecture to model short- and long-term temporal dependencies between frames, ultimately producing the final anomaly score and class prediction probabilities. We compare AnomalyCLIP against state-of-the-art methods considering three major anomaly detection benchmarks, i.e. ShanghaiTech, UCF-Crime, and XD-Violence, and empirically show that it outperforms baselines in recognising video anomalies. Project website and code are available at https://lucazanella.github.io/AnomalyCLIP/.

Feature boosting with efficient attention for scene parsing

The complexity of scene parsing grows with the number of object and scene classes, which is higher in unrestricted open scenes. The biggest challenge is to model the spatial relation between scene elements while succeeding in identifying objects at smaller scales. This paper presents a novel feature-boosting network that gathers spatial context from multiple levels of feature extraction and computes the attention weights for each level of representation to generate the final class labels. A novel ‘channel attention module’ is designed to compute the attention weights, ensuring that features from the relevant extraction stages are boosted while the others are attenuated. The model also learns spatial context information at low resolution to preserve the abstract spatial relationships among scene elements and reduce computational cost. Spatial attention is subsequently concatenated into a final feature set before applying feature boosting. Low-resolution spatial attention features are trained using an auxiliary task that help to learn a coarse global scene structure. The proposed model outperforms all state-of-the-art models on both the ADE20K and the Cityscapes datasets.

Achieving Optimal Order in a Novel Family of Numerical Methods: Insights from Convergence and Dynamical Analysis Results

In this manuscript, we introduce a novel parametric family of multistep iterative methods designed to solve nonlinear equations. This family is derived from a damped Newton’s scheme but includes an additional Newton step with a weight function and a “frozen” derivative, that is, the same derivative than in the previous step. Initially, we develop a quad-parametric class with a first-order convergence rate. Subsequently, by restricting one of its parameters, we accelerate the convergence to achieve a third-order uni-parametric family. We thoroughly investigate the convergence properties of this final class of iterative methods, assess its stability through dynamical tools, and evaluate its performance on a set of test problems. We conclude that there exists one optimal fourth-order member of this class, in the sense of Kung–Traub’s conjecture. Our analysis includes stability surfaces and dynamical planes, revealing the intricate nature of this family. Notably, our exploration of stability surfaces enables the identification of specific family members suitable for scalar functions with a challenging convergence behavior, as they may exhibit periodical orbits and fixed points with attracting behavior in their corresponding dynamical planes. Furthermore, our dynamical study finds members of the family of iterative methods with exceptional stability. This property allows us to converge to the solution of practical problem-solving applications even from initial estimations very far from the solution. We confirm our findings with various numerical tests, demonstrating the efficiency and reliability of the presented family of iterative methods.

Microbiological Assessment of Smartphone Surfaces Obtained from Final Year Students at Hezekiah University Umudi, Nkwerre, Imo State, Nigeria.

Smartphones of students in tertiary institutions have been described as an understudied and under-explored microcosm by leading public health experts. The microbial quality of smartphone surfaces has been a constant reason for several scientific discussions. Two hundred students were initially targeted for the study, ten students in the final year class at the Hezekiah University Umudi permitted access to their phones during the sampling. The study carried by swabbing the surface of their phones with a normal saline-moistened swab stick using a square quadrant. The content was dislodged and diluted using a 10-fold serial dilution technique. The samples were evaluated for total heterotrophic count, coliform, fungal and Staphylococcal count. The bacterial isolates were tentatively identified using biochemical reactions. The total heterotrophic count of the samples obtained from the male students ranged from Log 2.82 CFU/cm2 to Log 3.21 CFU/cm2 while the total coliform count was observed to range from Log 1.3 CFU/cm2 to Log 1.6 CFU/cm2. The total heterotrophic count ranged from Log 2.82 CFU/cm2 to Log 3.21 CFU/cm2 for male students and from Log 1.6 CFU/cm2 to Log 2.73 CFU/cm2 for female students. The isolates obtained from the study were Acinetobacter sp., Escherichia coli, Proteus sp., Klebsiella sp., Pseudomonas sp. and Staphylococcus sp. There is a need for the University to educate students on the need to routinely use alcohol-based wipes to clean the surfaces of their phones as a precautionary measure to limit diseases associated with the use of mobile phones.

Adversarially attack feature similarity for fine-grained visual classification

Fine-grained visual classification (FGVC) strives to distinguish images from distinct sub-classes within the same overarching meta-class, which is significant in various practical applications. Existing works mainly employ attention mechanisms to learn discriminative feature representations of objects under weakly supervised learning. In this paper, we argue that this likehood-based attention learning manner often outputs an inadequate feature representation since the available image-level labels fail to provide an explicit supervisory signal for attention learning, especially when the fine-grained images share a small and inconsistent inter-class variance. To alleviate this issue, we consider this challenging task from the perspective of attacking the feature representation between similar sub-classes to maximize the feature discriminativeness via learning adversarial examples, and propose an Adversarial-Aware Fine-Grained Visual Classification Network (A2Net). Specifically, we first propose an adversarial attack module based on projected gradient descent, which appends multiple-scale adversarial perturbations to simulate subclass examples with different similarities. Then, we introduce an adversarial attention generation module that estimates the effect of attention learned on adversarial examples and legitimate examples for the final class prediction through causal inference. The adversarial attention generation module is encouraged to maximize the effect, which provides powerful supervision to capture more attention indicating the discriminative parts. We further propose an adversarial-aware module to learn the feature-level differences between legitimate and adversarial examples, which helps enhance the semantic boundaries of class-specific features for accurate FGVC. The extensive array of experiments conducted serves to underscore the efficacy of the proposed A2Net, outperforming state-of-the-art FGVC methods on CUB-200–2011, FGVC-Aircraft, Stanford Cars, Stanford Dogs, and NABirds benchmarks.

Dampak Ketergantungan Konten Flexing pada Fase Quarter Life Crisis Mahasiswa Tingkat Akhir

Final-year students aged 20 years and over often feel various pressures to achieve a certain level of success or achievement. One of them is the pressure from flexing content that is often presented on social media. Dependence on flexing content can trigger anxiety about the future, which ultimately leads to the emergence of the quarter-life crisis phase. Through its content, social media indirectly influences students, especially in forming a crisis of self-confidence. This research aims to determine the influence of dependency on flexing content on the quarter-life crisis phase of final-year students and to find out how big the influence of dependency on flexing content is on the quarter-life crisis phase of final-year students. The media dependency theory introduced by Rokeach & Defleur (1976) is utilized as the primary foundation for this research. This research uses a quantitative approach with a survey method for 214 Communication Science students class of 2020 at Telkom University, Bandung. The results of this research obtained a significance value of 0.000 <0.05. The research findings indicate that the dependency on flexing content on social media has a considerable potential to induce anxiety about life direction, leading to the formation of the quarter-life crisis phase. This shows that the dependence on flexing content which includes cognitive, affective, and behavioral aspects influences the quarter-life crisis phase which includes hopes and dreams, educational challenges, religion and spirituality, work life, home, lovers, family, and friends, and identity in final-year students. final Communication Science class of 2020, Telkom University Bandung, with an influence level of 40.7%, and the other 59.3% were influenced by other factors not included in the research. Research implications and suggestions are also discussed.

Incorporating ecological economics into engineering education

Engineering students require an education which facilitates the development not just of functional design of technological artefacts and processes, but they also need to be equipped to understand and engage with their wider implications and context. Consequently, as part of the formation of sustainability informed engineers, there is a clear necessity to integrate the socio-economic aspects of sustainability in order to posit technology in appropriate contextualised settings. The chemical engineering degree programme at University College Cork incorporates a strong emphasis on integrating the socio-economic dimension of sustainability within the programme. This paper focuses on ecological economics, and how it can connect with engineering and why it is important that engineering students receive a grounding in ecological economics. It outlines the content of a five-lecture series on ecological economics given to fifth-year chemical engineering students. This includes environmental aspects, such ecological limits analysis and the use of strategies, including economic, to move unsustainable natural resource use and emissions discharges to within their ecological limits. It explores social aspects and the critical barriers to transitioning to a sustainable economy, as well as the impact of ecological economics on engineering. Finally, three consecutive final year classes were surveyed to elicit their feedback on being taught ecological economics and to test the hypothesis that the inclusion of ecological economics appreciably enhances students’ ability to engage productively with sustainability and the UN Sustainable Development Goals (SDGs); a key accreditation and professional requirement for contemporary engineering graduates. Overall, most students enjoyed the material, stating that is of value to engineering students, and they provided a useful qualitative assessment of the content.

Classification of synoptic circulation patterns with a two-stage clustering algorithm using the modified structural similarity index metric (SSIM)

Abstract. We develop a new classification method for synoptic circulation patterns with the aim to extend the evaluation routine for climate simulations. This classification is applicable to any region of the globe of any size given the reference data. Its unique novelty is the use of the modified structural similarity index metric (SSIM) instead of traditional distance metrics for cluster building. This classification method combines two classical clustering algorithms used iteratively, hierarchical agglomerative clustering (HAC) and k-medoids, with only one pre-set parameter – the threshold on the similarity between two synoptic patterns expressed as the structural similarity index measure (SSIM). This threshold is set by the user to imitate the human perception of the similarity between two images (similar structure, luminance, and contrast), whereby the number of final classes is defined automatically. We apply the SSIM-based classification method to the geopotential height at the pressure level of 500 hPa from the ERA-Interim reanalysis data for 1979–2018 and demonstrate that the built classes are (1) consistent with the changes in the input parameter, (2) well-separated, (3) spatially stable, (4) temporally stable, and (5) physically meaningful. We demonstrate an exemplary application of the synoptic circulation classes obtained with the new classification method for evaluating Coupled Model Intercomparison Project Phase 6 (CMIP6) historical climate simulations and an alternative reanalysis (for comparison purposes): output fields of CMIP6 simulations (and of the alternative reanalysis) are assigned to the classes and the Jensen–Shannon distance is computed for the match in frequency, transition, and duration probabilities of these classes. We propose using this distance metric to supplement a set of commonly used metrics for model evaluation.

Entrepreneurial education in a pandemic era: Timeframes, demographics, and the nexus between teaching and experiential learning

Purpose: This study examines the transformative effects of the COVID-19 pandemic on entrepreneurial education during and after the pandemic. Methodology: Using action research-based case study with a mixed data collection method, that spans pre-, during, and post-COVID-19. Data collection involved class discussions in a final mandatory class, surveys sent after the end of the class, and official course evaluations. Findings: The pandemic-induced changes highlighted that extended instructional timeframes do not necessarily yield better entrepreneurial outcomes. A condensed, immersive approach fostered innovation and improved entrepreneurial mindsets. The hybrid teaching model introduced by the pandemic enriched class demographics, offering a broader spectrum of perspectives and enhancing discussions. Implications: The shift to digital and hybrid teaching models brought diverse student profiles, which, when appropriately harnessed, can enrich classroom discussions and peer learning. Future research should delve deeper into the motivations of distinct student groups and the long-term effects of current teaching strategies on entrepreneurial endeavors.

Sharp-SSL: Selective high-dimensional axis-aligned random projections for semi-supervised learning

We propose a new method for high-dimensional semi-supervised learning problems based on the careful aggregation of the results of a low-dimensional procedure applied to many axis-aligned random projections of the data. Our primary goal is to identify important variables for distinguishing between the classes; existing low-dimensional methods can then be applied for final class assignment. To this end, we score projections according to their class-distinguishing ability; for instance, motivated by a generalized Rayleigh quotient, we can compute the traces of estimated whitened between-class covariance matrices on the projected data. This enables us to assign an importance weight to each variable for a given projection, and to select our signal variables by aggregating these weights over high-scoring projections. Our theory shows that the resulting Sharp-SSL algorithm is able to recover the signal coordinates with high probability when we aggregate over sufficiently many random projections and when the base procedure estimates the diagonal entries of the whitened between-class covariance matrix sufficiently well. For the Gaussian EM base procedure, we provide a new analysis of its performance in semi-supervised settings that controls the parameter estimation error in terms of the proportion of labeled data in the sample. Numerical results on both simulated data and a real colon tumor dataset support the excellent empirical performance of the method.

Teaching/Learning Philosophy In High School: Case Of Congo

This article is a study and research carried out among students in final year literary classes and Philosophy teachers, at the Nganga Edouard high school in Brazzaville Congo), during the 2019-2020 school year. We wanted to answer the following question: what are the didactic and pedagogical problems that philosophy teachers encounter in the transmission of knowledge related to existence and its problems? This question is linked to the didactics and pedagogy of philosophy in high school. Propose didactics and pedagogy of philosophy, more adapted to the theme entitled “the philosophical teaching of existence and its problems”. We arrive at the following results: the difficulties of philosophical teaching on existence and its problems in literary final year classes are linked to the misuse of didactics and pedagogy by teachers; the theme of existence and its problems is not well taught in high school because of the limitations experienced by teachers regarding the methodical explanation of the subjects; the teaching of this theme is too theoretical and far from the real existence and daily experience of learners. We suggest an interactive didactic approach; a participatory pedagogy; an integration of the real problems of the learners' existence, in the content of the lessons.

Peer Critiquing in Scientific Inquiry: Enhancing Preservice Teachers' Reasoning, Epistemic Knowledge, and Scientific Argumentation through Innovative Teaching Strategies in an Undergraduate Science Content Course

The study explored the impact of peer critiquing on preservice teachers' understanding of the nature of science, scientific inquiry, and argumentation in an undergraduate science content course. The aim was to investigate innovative teaching strategies that enhance preservice teachers' comprehension of these key aspects. This study showcased implementing these strategies in which the preservice teachers actively critique and evaluate each other's work, fostering a culture of constructive feedback and reflection. Data were collected from 38 preservice teachers from three data sources that included a group argumentation of an assigned environment scenario of the endangered species, a peer critique of a video presentation of the scientific explanation of an individual environmental science topic, and finally a peer critique on someone else’s scientific inquiry study that could either be observational or experimental. To achieve this goal, we used the class activities to incorporate new ideas within the classroom by documenting the preservice teachers’ practices as they reflect upon their peers’ practices. While video critiques indicated a general competency, there was a need for further development in analyzing scientific concepts. The final class critiquing session demonstrated preservice teachers' self-awareness and understanding of the importance of relevant research in scientific inquiry. Peer critiquing activities positively impacted preservice teachers' reasoning skills, as evidenced by their ability to construct well-reasoned arguments and analyze empirical evidence. By examining preservice teachers' experiences and outcomes, we seek to provide valuable insights and recommendations for educators striving to enhance preservice teachers' understanding of the nature of science and scientific inquiry through peer critiquing.

인공지능 리터러시를 위한 AI 융합 영어 쓰기 수업 모형 개발

Objectives This study aimed to develop a model for an English writing class integrating Generative AI for artificial intelligence literacy.
 Methods Literature and prior research on artificial intelligence literacy, prompt engineering, Generative AI, and the development of English writing class models were analyzed. The model prototype's purpose and theoretical foundation were established, validated for internal consistency by five experts in educational technology and English education. The prototype was tested with one English teacher and approximately 48 students over a two-week period in December 2023, refining its content for the final model.
 Results To achieve this, 31 first-year middle school students in Seoul were assessed through pre- and post-tests to measure their attitudes toward artificial intelligence literacy and English writing. The final class model consisted of seven stages, distinguishing between learner, instructor, and Generative AI roles. Nine teaching strategies were applied to 39 detailed stages of the class model, aligning with instructional design principles.
 Conclusions The developed model holds significance in encouraging students' divergent questioning and proactive learning attitudes through strategic and systematic integration of Generative AI interaction within the English writing class. Moreover, through the development of this class model and teaching strategies, it is anticipated to serve as an effective guideline for educators to design and implement prompt engineering-based Generative AI integrated English writing classes.

Hubungan Penggunaan Gadget Dan Stres Dengan Insomnia Pada Remaja Kelas Akhir Di SMK Kesehatan Logos Kabupaten Bogor Tahun 2023

Sleep is something that must be fulfilled, if it it fulfilled it will cause disturbances in a person's biology and psychology. Many factors influence insomnia, including disease, fatigue, environment, stress, medication, diet, and alcohol and nicotine. The use of gadgets can also affect insomnia because addiction to bluelight gadgets will shift your sleep rhythm, causing insomnia. This research aims to determine the relationship between intensity of gadget and stress with insomnia in final class adolescents at Logos Health Vocational School. The research method used is Cross-Sectional. involving 99 final grade students of Logos Health Vocational School. By using the Chi-Square Statistical Test and the results of this research, the P value = 0.001, which means that the value is <0.05. Conclusion: There is a relationship between the intensity of gadget and stress and insomnia in final class adolescents at Logos Health Vocational School, Kabupaten Bogor in 2023. The results of this research are expected to provide benefits for parents and schools in efforts to prevent the occurrence of insomnia, as well as handling stress and use. gadgets.

Effects of using audience response systems (Kahoot and Google Forms and Sheets) on learning of medical students in a large class: A questionnaire-based retrospective study

AbstractThe effectiveness of implementing various audience response system (ARS) formats in large classes to engage students, positively influence academic outcomes, and align with their preferences for different types of activities lacked sufficient information. This study aimed to (1) evaluate students’ perspectives on different ARS formats, including Kahoot for individual (Kahoot-I) and team (Kahoot-T) play, and Google Forms and Google Sheets for multiple-choice questions (Google-MCQ) and for open-ended questions (Google-OEQ) on different aspects with/without subgroup analysis of students into quartiles (Q1-lowest-Q4-highest) of summative scores, and (2) determine correlations between students’ perspectives and academic outcomes. At the course’s final class, a QR code for a 5-point Likert scale questionnaire was displayed for all enrolled medical students. 269/312 students (86.22%) provided written consent and completed the questionnaire constructed based on dimensions of student engagement. Across all ARS formats, students rated scores ≥ 4 for almost all aspects, with participation in class receiving the highest scores. For each aspect, Kahoot-T was rated highest in various aspects and Google-MCQ for “understanding” and Google-OEQ for “asking/answering question”. Higher academic achievers expressed more “enthusiasm” for Kahoot-I than lower achievers (P < 0.05). Formative and summative scores were positively correlated with many aspects. Students reported high engagement with all ARS formats, with higher academic achievers rating them with higher scores. Kahoot-T was the most favored for multiple aspects, Google-MCQ for “understanding” and Google-OEQ for interactive “asking/answering question”. Thus, selecting ARS in alignment with learning objective could enhance students’ engagement and learning outcomes.

APPLICATION OF THE EDUCATIONAL INFORMATION AND COMMUNICATION TECHNOLOGIES FOR TEACHING FOREIGN LANGUAGES TO STUDENTS MAJORING IN INFORMATION TECHNOLOGIES

The article presents the overview of cutting-edge educational information and communication technologies used by the lecturers of the Department of Cross- Cultural Professional Communication of the Faculty of Engineering and Economics of the Belarusian State University of Informatics and Radioelectronics for teaching foreign languages to the undergraduate students of the first stage of higher education receiving a bachelor’s degree in information technologies. The article also analyses the effectiveness of the Telegram-based information channels used by the lecturers of foreign languages as an alternative to sending materials to students via e-mail. The author outlines an interactive web-based resource slido.com for visualising the results of brainstorming sessions instead of using the traditional ways of presenting the results of this teamwork format. The article describes a non-standard way of introducing new vocabulary with the help of rebuses created on a web-based resource rebus.club which does not require any prior registration. The author also presents an additional way of assessment of the level of proficiency in oral examination topics and revision of their content within the final classes after completing the study of a module or section in the format of an interactive Jeopardy game created with the help of a template based on a web-based resource jeopardylabs.com. To conduct current and final lexical and grammar tests on the topics covered within the course of practical classes of a foreign language, as well as to allow students to evaluate their success in the classroom, a web-based resource onlinetestpad.com is offered to replace traditional methods of assessment of the undergraduate students of the first stage of higher education receiving a bachelor’s degree in information technologies.

Comparison of CNN-based methods for yoga pose classification

Yoga is an exercise developed in ancient India. People perform yoga in order to have mental, physical, and spiritual benefits. While yoga helps build strength in the mind and body, incorrect postures might result in serious injuries. Therefore, yoga exercisers need either an expert or a platform to receive feedback on their performance. Since access to experts is not an option for everyone, a system to provide feedback on the yoga poses is required. To this end, commercial products such as smart yoga mats and smart pants are produced; Kinect cameras, sensors, and wearable devices are used. However, these solutions are either uncomfortable to wear or not affordable for everyone. Nonetheless, a system that employs computer vision techniques is a requirement. In this paper, we propose a deep-learning model for yoga pose classification, which is the first step of a quality assessment and personalized feedback system. We introduce a wavelet-based model that first takes wavelet transform of input images. The acquired subbands, i.e., approximation, horizontal, vertical, and diagonal coefficients of the wavelet transform are then fed into separate convolutional neural networks (CNN). The obtained probability results for each group are fused to predict the final yoga class. A publicly available dataset with 5 yoga poses is used. Since the number of images in the dataset is not enough for a deep learning model, we also perform data augmentation to increase the number of images. We compare our results to a CNN model and the three models that employ the subbands separately. Results obtained using the proposed model outperforms the accuracy output achieved with the compared models. While the regular CNN model has 61% and 50% accuracy for the training and test data, the proposed model achieves 91% and 80%, respectively.