Abdelrahman Halawa, S. Gamalel-Din, Abdurrahman A. Nasr
Writing a well-structured scientific documents, such as articles and theses, is vital for comprehending the document's argumentation and understanding its messages. Furthermore, it has an impact on the efficiency and time required for studying the document. Proper document segmentation also yields better results when employing automated Natural Language Processing (NLP) manipulation algorithms, including summarization and other information retrieval and analysis functions. Unfortunately, inexperienced writers, such as young researchers and graduate students, often struggle to produce well-structured professional documents. Their writing frequently exhibits improper segmentations or lacks semantically coherent segments, a phenomenon referred to as "mal-segmentation." Examples of mal-segmentation include improper paragraph or section divisions and unsmooth transitions between sentences and paragraphs. This research addresses the issue of mal-segmentation in scientific writing by introducing an automated method for detecting mal-segmentations, and utilizing Sentence Bidirectional Encoder Representations from Transformers (sBERT) as an encoding mechanism. The experimental results section shows a promising results for the detection of mal-segmentation using the sBERT technique.
{"title":"EXPLOITING BERT FOR MALFORMED SEGMENTATION DETECTION TO IMPROVE SCIENTIFIC WRITINGS","authors":"Abdelrahman Halawa, S. Gamalel-Din, Abdurrahman A. Nasr","doi":"10.35784/acs-2023-20","DOIUrl":"https://doi.org/10.35784/acs-2023-20","url":null,"abstract":"Writing a well-structured scientific documents, such as articles and theses, is vital for comprehending the document's argumentation and understanding its messages. Furthermore, it has an impact on the efficiency and time required for studying the document. Proper document segmentation also yields better results when employing automated Natural Language Processing (NLP) manipulation algorithms, including summarization and other information retrieval and analysis functions. Unfortunately, inexperienced writers, such as young researchers and graduate students, often struggle to produce well-structured professional documents. Their writing frequently exhibits improper segmentations or lacks semantically coherent segments, a phenomenon referred to as \"mal-segmentation.\" Examples of mal-segmentation include improper paragraph or section divisions and unsmooth transitions between sentences and paragraphs. This research addresses the issue of mal-segmentation in scientific writing by introducing an automated method for detecting mal-segmentations, and utilizing Sentence Bidirectional Encoder Representations from Transformers (sBERT) as an encoding mechanism. The experimental results section shows a promising results for the detection of mal-segmentation using the sBERT technique.","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41933803","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Classification plays a critical role in machine learning (ML) systems for processing images, text and high -dimensional data. Predicting class labels from training data is the primary goal of classification. An optimal model for a particular classification problem is chosen on the basis of the model's performance and execution time. This paper compares and analyses the performance of basic as well as ensemble classifiers utilizing 10 -fold cross validation and also discusses their essential concepts, advantages, and disadvantages. In this study five basic classifiers namely Naïve Bayes (NB), Multi-layer Perceptron (MLP), Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF) and the ensemble of all the five classifiers along with few more combinations are compared with five University of California Irvine (UCI) ML Repository datasets and a Diabetes Health Indicators dataset from kaggle repository. To analyze and compare the performance of classifiers, evaluation metrics like Accuracy, Recall, Precision, Area Under Curve (AUC) and F-Score are used. Experimental results showed that SVM performs best on two out of the six datasets (Diabetes Health Indicators and waveform), RF performs best for Arrhythmia, Sonar, Tic-tac-toe datasets, and the best ensemble combination is found to be DT+SVM+RF on Ionosphere dataset having respective accuracies 72.58%, 90.38%, 81.63%, 73.59%, 94.78% and 94.01% and the proposed ensemble combinations outperformed over the conventional models for few datasets.
{"title":"A COMPARATIVE STUDY ON PERFORMANCE OF BASIC AND ENSEMBLE CLASSIFIERS WITH VARIOUS DATASETS","authors":"Archana Gunakala, Afzal Hussain Shahid","doi":"10.35784/acs-2023-08","DOIUrl":"https://doi.org/10.35784/acs-2023-08","url":null,"abstract":"Classification plays a critical role in machine learning (ML) systems for processing images, text and high -dimensional data. Predicting class labels from training data is the primary goal of classification. An optimal model for a particular classification problem is chosen on the basis of the model's performance and execution time. This paper compares and analyses the performance of basic as well as ensemble classifiers utilizing 10 -fold cross validation and also discusses their essential concepts, advantages, and disadvantages. In this study five basic classifiers namely Naïve Bayes (NB), Multi-layer Perceptron (MLP), Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF) and the ensemble of all the five classifiers along with few more combinations are compared with five University of California Irvine (UCI) ML Repository datasets and a Diabetes Health Indicators dataset from kaggle repository. To analyze and compare the performance of classifiers, evaluation metrics like Accuracy, Recall, Precision, Area Under Curve (AUC) and F-Score are used. Experimental results showed that SVM performs best on two out of the six datasets (Diabetes Health Indicators and waveform), RF performs best for Arrhythmia, Sonar, Tic-tac-toe datasets, and the best ensemble combination is found to be DT+SVM+RF on Ionosphere dataset having respective accuracies 72.58%, 90.38%, 81.63%, 73.59%, 94.78% and 94.01% and the proposed ensemble combinations outperformed over the conventional models for few datasets.","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45767243","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The quality dimensions of an information system, such as system, information, and service qualities, play a crucial role in determining the overall performance of an organization. These quality dimensions are significant as they can impact employee outcomes, which are key factors in determining whether an organization is able to achieve a competitive advantage in the market. The aim of this study is to explore the impact of quality dimensions on employee outcomes such as learning ability, adaptability, and job satisfaction. The research was conducted by distributing a structured survey questionnaire to 300 employees of 8 commercial banks at different management levels. The measurement and structural models were analyzed using Smart PLS. This study employed descriptive analysis to present a comprehensive demographic profile of both the organizations and the participants. Out of the nine hypotheses tested, seven were found to be significant. The findings of this study show that while all three quality dimensions (system, information, and service) of information systems positively affect employee learning, only system and information qualities positively affect employee learning, and as for job satisfaction, only system and service qualities play an important role. Therefore, implementing suitable information systems to improve employee outcomes in an organization, especially a financial organization, is paramount in this information age. This research contributes to understanding information systems, their implementation, and employee outcomes in an organization.
{"title":"CAN THE SYSTEM, INFORMATION, AND SERVICE QUALITIES IMPACT EMPLOYEE LEARNING, ADAPTABILITY, AND JOB SATISFACTION?","authors":"Zahid B. Zamir","doi":"10.35784/acs-2023-03","DOIUrl":"https://doi.org/10.35784/acs-2023-03","url":null,"abstract":"The quality dimensions of an information system, such as system, information, and service qualities, play a crucial role in determining the overall performance of an organization. These quality dimensions are significant as they can impact employee outcomes, which are key factors in determining whether an organization is able to achieve a competitive advantage in the market. The aim of this study is to explore the impact of quality dimensions on employee outcomes such as learning ability, adaptability, and job satisfaction. The research was conducted by distributing a structured survey questionnaire to 300 employees of 8 commercial banks at different management levels. The measurement and structural models were analyzed using Smart PLS. This study employed descriptive analysis to present a comprehensive demographic profile of both the organizations and the participants. Out of the nine hypotheses tested, seven were found to be significant. The findings of this study show that while all three quality dimensions (system, information, and service) of information systems positively affect employee learning, only system and information qualities positively affect employee learning, and as for job satisfaction, only system and service qualities play an important role. Therefore, implementing suitable information systems to improve employee outcomes in an organization, especially a financial organization, is paramount in this information age. This research contributes to understanding information systems, their implementation, and employee outcomes in an organization.","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44771814","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Abderrahim Bahani, Elhoussine Ech-Chhibat, H. Samri, Laila AIT MAALEM, Hicham AIT EL ATTAR
This paper presents a Multiple Adaptive Neuro-Fuzzy Inference System (MANFIS)-based method for regulating the handling force of a common object. The foundation of this method is the prediction of the inverse dynamics of a cooperative robotic system made up of two 3-DOF robotic manipulators. Considering the no slip in contact between the tool and the object, an object is moved. to create and feed the MANFIS database, the inverse kinematics and dynamic equations of motion for the closed chain of motion for both arms are established in Matlab. Results from a SimMechanic simulation are given to demonstrate how well the suggested ANFIS controller works. Several manipulated object movements covering the shared workspace of the two manipulator arms are used to test the proposed control strategy.
{"title":"INTELLIGENT CONTROLLING THE GRIPPING FORCE OF AN OBJECT BY TWO COMPUTER-CONTROLLED COOPERATIVE ROBOTS","authors":"Abderrahim Bahani, Elhoussine Ech-Chhibat, H. Samri, Laila AIT MAALEM, Hicham AIT EL ATTAR","doi":"10.35784/acs-2023-09","DOIUrl":"https://doi.org/10.35784/acs-2023-09","url":null,"abstract":"This paper presents a Multiple Adaptive Neuro-Fuzzy Inference System (MANFIS)-based method for regulating the handling force of a common object. The foundation of this method is the prediction of the inverse dynamics of a cooperative robotic system made up of two 3-DOF robotic manipulators. Considering the no slip in contact between the tool and the object, an object is moved. to create and feed the MANFIS database, the inverse kinematics and dynamic equations of motion for the closed chain of motion for both arms are established in Matlab. Results from a SimMechanic simulation are given to demonstrate how well the suggested ANFIS controller works. Several manipulated object movements covering the shared workspace of the two manipulator arms are used to test the proposed control strategy.","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42726876","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the paper, the authors are presenting the analysis of implementation of IoT system of road quality analysis. The proposed system has been prepared with edge, on-device processing in mind, allowing for reduction of amount of data being sent to cloud computing aggregation subsystem, sending only 2.5% of the original data. Several algorithms for road quality analysis has been implemented on a real device and tested in a real-world conditions. The system has been compared to the state-of-the-art offline processing approach and shown very similar results.
{"title":"USAGE OF IOT EDGE APPROACH FOR ROAD QUALITY ANALYSIS","authors":"M. Badurowicz, Sebastian Łagowski","doi":"10.35784/acs-2023-02","DOIUrl":"https://doi.org/10.35784/acs-2023-02","url":null,"abstract":"In the paper, the authors are presenting the analysis of implementation of IoT system of road quality analysis. The proposed system has been prepared with edge, on-device processing in mind, allowing for reduction of amount of data being sent to cloud computing aggregation subsystem, sending only 2.5% of the original data. Several algorithms for road quality analysis has been implemented on a real device and tested in a real-world conditions. The system has been compared to the state-of-the-art offline processing approach and shown very similar results.","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46327511","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We present a machine learning predictor for academic results datasets (PARD), for missing academic results based on chi-squared expected calculation, positional clustering, progressive approximation of relative residuals, and positional averages of the data in a sampled population. Academic results datasets are data originating from academic institutions’ results repositories. It is a technique designed specifically for predicting missing academic results. Since the whole essence of data mining is to elicit useful information and gain knowledge-driven insights into datasets, PARD positions data explorer at this advantageous perspective. PARD promises to solve missing academic results dataset problems more quickly over and above what currently obtains in literatures. The predictor was implemented using Python, and the results obtained show that it is admissible in a minimum of up to 93.6 average percent accurate predictions of the sampled cases. The results demonstrate that PARD shows a tendency toward greater precision in providing the better solution to the problems of predictions of missing academic results datasets in universities.
{"title":"ARDP: SIMPLIFIED MACHINE LEARNING PREDICTOR FOR MISSING UNIDIMENSIONAL ACADEMIC RESULTS DATASET","authors":"O. Folorunso, O. Akinyede, K. Agbele","doi":"10.35784/acs-2023-04","DOIUrl":"https://doi.org/10.35784/acs-2023-04","url":null,"abstract":"We present a machine learning predictor for academic results datasets (PARD), for missing academic results based on chi-squared expected calculation, positional clustering, progressive approximation of relative residuals, and positional averages of the data in a sampled population. Academic results datasets are data originating from academic institutions’ results repositories. It is a technique designed specifically for predicting missing academic results. Since the whole essence of data mining is to elicit useful information and gain knowledge-driven insights into datasets, PARD positions data explorer at this advantageous perspective. PARD promises to solve missing academic results dataset problems more quickly over and above what currently obtains in literatures. The predictor was implemented using Python, and the results obtained show that it is admissible in a minimum of up to 93.6 average percent accurate predictions of the sampled cases. The results demonstrate that PARD shows a tendency toward greater precision in providing the better solution to the problems of predictions of missing academic results datasets in universities.","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41619325","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
As the basic technology of human action recognition, pose estimation is attracting more and more researchers' attention, while edge application scenarios pose a higher challenge. This paper proposes a lightweight multi-person pose estimation scheme to meet the needs of real-time human action recognition on the edge end. This scheme uses AlphaPose to extract human skeleton nodes, and adds ResNet and Dense Upsampling Revolution to improve its accuracy. Meanwhile, we use YOLO to enhance AlphaPose’s support for multi-person pose estimation, and optimize the proposed model with TensorRT. In addition, this paper sets Jetson Nano as the Edge AI deployment device of the proposed model and successfully realizes the model migration to the edge end. The experimental results show that the speed of the optimized object detection model can reach 20 FPS, and the optimized multi-person pose estimation model can reach 10 FPS. With the image resolution of 320×240, the model’s accuracy is 73.2%, which can meet the real-time requirements. In short, our scheme can provide a basis for lightweight multi-person action recognition scheme on the edge end.
{"title":"A LIGHTWEIGHT MULTI-PERSON POSE ESTIMATION SCHEME BASED ON JETSON NANO","authors":"Lei Liu, E. Blancaflor, Mideth B. Abisado","doi":"10.35784/acs-2023-01","DOIUrl":"https://doi.org/10.35784/acs-2023-01","url":null,"abstract":"As the basic technology of human action recognition, pose estimation is attracting more and more researchers' attention, while edge application scenarios pose a higher challenge. This paper proposes a lightweight multi-person pose estimation scheme to meet the needs of real-time human action recognition on the edge end. This scheme uses AlphaPose to extract human skeleton nodes, and adds ResNet and Dense Upsampling Revolution to improve its accuracy. Meanwhile, we use YOLO to enhance AlphaPose’s support for multi-person pose estimation, and optimize the proposed model with TensorRT. In addition, this paper sets Jetson Nano as the Edge AI deployment device of the proposed model and successfully realizes the model migration to the edge end. The experimental results show that the speed of the optimized object detection model can reach 20 FPS, and the optimized multi-person pose estimation model can reach 10 FPS. With the image resolution of 320×240, the model’s accuracy is 73.2%, which can meet the real-time requirements. In short, our scheme can provide a basis for lightweight multi-person action recognition scheme on the edge end.","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45830347","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Material flow management aims to ensure the consistency of supply and reliability of the production processes being carried out. The aim of the article is to present a model of material flow organisation in a changing production system operating under small batch production conditions. Carrying out simulations for various production scenarios will be the basis for developing an effective method of material flow management in small batch production of cutting tools.
{"title":"IMPROVING MATERIAL FLOW IN A MODIFIED PRODUCTION SYSTEM","authors":"D. Plinta, K. Radwan","doi":"10.35784/acs-2023-07","DOIUrl":"https://doi.org/10.35784/acs-2023-07","url":null,"abstract":"Material flow management aims to ensure the consistency of supply and reliability of the production processes being carried out. The aim of the article is to present a model of material flow organisation in a changing production system operating under small batch production conditions. Carrying out simulations for various production scenarios will be the basis for developing an effective method of material flow management in small batch production of cutting tools.","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49371022","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The Internet of Things (IoT) touches almost every aspect of modern society and has changed the way people live, work, travel and, do business. Because of its importance, it is essential to ensure that an IoT system is performing well, as desired and expected, and that this can be assessed and managed with an adequate set of IoT performance metrics. The aim of this study was to systematically inventory and classifies recent studies that have investigated IoT metrics. We conducted a literature review based on studies published between January 2010 and December 2021 using a set of five research questions (RQs) on the current knowledge bases for IoT metrics. A total of 158 IoT metrics were identified and classified into 12 categories according to the different parts and aspects of an IoT system. To cover the overall performance of an IoT system, the 12 categories were organized into an ontology. The findings results show that the category of network metrics was the most discussed in 43% of the studies and, with the highest number of metrics at 37%. This study can provide guidelines for researchers and practitioners in selecting metrics for IoT systems and valuable insights into areas for improvement and optimization.
{"title":"SYSTEMATIC LITERATURE REVIEW OF IOT METRICS","authors":"Donatien Koulla Moulla, Ernest Mnkandla, A. Abran","doi":"10.35784/acs-2023-05","DOIUrl":"https://doi.org/10.35784/acs-2023-05","url":null,"abstract":"The Internet of Things (IoT) touches almost every aspect of modern society and has changed the way people live, work, travel and, do business. Because of its importance, it is essential to ensure that an IoT system is performing well, as desired and expected, and that this can be assessed and managed with an adequate set of IoT performance metrics. The aim of this study was to systematically inventory and classifies recent studies that have investigated IoT metrics. We conducted a literature review based on studies published between January 2010 and December 2021 using a set of five research questions (RQs) on the current knowledge bases for IoT metrics. A total of 158 IoT metrics were identified and classified into 12 categories according to the different parts and aspects of an IoT system. To cover the overall performance of an IoT system, the 12 categories were organized into an ontology. The findings results show that the category of network metrics was the most discussed in 43% of the studies and, with the highest number of metrics at 37%. This study can provide guidelines for researchers and practitioners in selecting metrics for IoT systems and valuable insights into areas for improvement and optimization. \u0000 ","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45429426","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Supervised learning as a sub-discipline of machine learning enables the recognition of correlations between input variables (features) and associated outputs (classes) and the application of these to previously unknown data sets. In addition to typical areas of application such as speech and image recognition, fields of applications are also being developed in the sports and fitness sector. The purpose of this work was to implement a workflow for the automated recognition of sports exercises in the Matlab® programming environment and to carry out a comparison of different model structures. First, the acquisition of the sensor signals provided in the local network and their processing were implemented. The functionalities to be realised included the interpolation of lossy time series, the labelling of the activity intervals performed and, in part, the generation of sliding windows with statistical parameters. The preprocessed data were used for the training of classifiers and artificial neural networks (ANN). These were iteratively optimised in their corresponding hyper parameters for the data structure to be learned. The most reliable models were finally trained with an increased data set, validated and compared with regard to the achieved performance. In addition to the usual evaluation metrics such as F1 score and accuracy, the temporal behaviour of the assignments was also displayed graphically, which enabled statements to be made about potential causes for incorrect assignments. In this context, especially the transition areas between the classes were detected as erroneous assignments as well as exercises with insufficient or clearly deviating execution. The best overall accuracy achieved with ANN and the increased dataset was 93.7 %.
{"title":"RECOGNITION OF SPORTS EXERCISES USING INERTIAL SENSOR TECHNOLOGY","authors":"P. Krutz, M. Rehm, H. Schlegel, Martin Dix","doi":"10.35784/acs-2023-10","DOIUrl":"https://doi.org/10.35784/acs-2023-10","url":null,"abstract":"Supervised learning as a sub-discipline of machine learning enables the recognition of correlations between input variables (features) and associated outputs (classes) and the application of these to previously unknown data sets. In addition to typical areas of application such as speech and image recognition, fields of applications are also being developed in the sports and fitness sector. The purpose of this work was to implement a workflow for the automated recognition of sports exercises in the Matlab® programming environment and to carry out a comparison of different model structures. First, the acquisition of the sensor signals provided in the local network and their processing were implemented. The functionalities to be realised included the interpolation of lossy time series, the labelling of the activity intervals performed and, in part, the generation of sliding windows with statistical parameters. The preprocessed data were used for the training of classifiers and artificial neural networks (ANN). These were iteratively optimised in their corresponding hyper parameters for the data structure to be learned. The most reliable models were finally trained with an increased data set, validated and compared with regard to the achieved performance. In addition to the usual evaluation metrics such as F1 score and accuracy, the temporal behaviour of the assignments was also displayed graphically, which enabled statements to be made about potential causes for incorrect assignments. In this context, especially the transition areas between the classes were detected as erroneous assignments as well as exercises with insufficient or clearly deviating execution. The best overall accuracy achieved with ANN and the increased dataset was 93.7 %.","PeriodicalId":36379,"journal":{"name":"Applied Computer Science","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43926875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：