Pub Date : 2023-02-23DOI: 10.1109/ECCE57851.2023.10101543
Md. Zubaer Alam, T. K. Roy, Subarto Kumar Ghosh, N. Mohammad, L. C. Paul
This research presents an improved backstepping control (IBSC) approach to designing a controller for a DC-DC buck converter to improve output voltage regulation under changing operating conditions. To develop the proposed the proposed controller, a state-space DC-DC buck converter dynamical model in continuous conduction mode is first developed. Secondly, to avoid the complexity of virtual control law derivatives in the traditional BSC method, these terms are treated as uncertain terms during the control law design process. Furthermore, the Lyapunov control theory is used to ensure the closed-loop system's global asymptotic stability. Finally, the performance of the proposed IBSC technique is validated using a simulation study on the MATLAB Simulink platform. A comparison of the simulation results is also presented to show the superiority of the proposed approach as compared to the traditional BSC method. The simulation study and quantitative results reveal that the proposed IBSC method outperforms the traditional BSC method.
{"title":"Output Voltage Stability of a DC-DC Buck Converter via an Improved Backstepping Method","authors":"Md. Zubaer Alam, T. K. Roy, Subarto Kumar Ghosh, N. Mohammad, L. C. Paul","doi":"10.1109/ECCE57851.2023.10101543","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101543","url":null,"abstract":"This research presents an improved backstepping control (IBSC) approach to designing a controller for a DC-DC buck converter to improve output voltage regulation under changing operating conditions. To develop the proposed the proposed controller, a state-space DC-DC buck converter dynamical model in continuous conduction mode is first developed. Secondly, to avoid the complexity of virtual control law derivatives in the traditional BSC method, these terms are treated as uncertain terms during the control law design process. Furthermore, the Lyapunov control theory is used to ensure the closed-loop system's global asymptotic stability. Finally, the performance of the proposed IBSC technique is validated using a simulation study on the MATLAB Simulink platform. A comparison of the simulation results is also presented to show the superiority of the proposed approach as compared to the traditional BSC method. The simulation study and quantitative results reveal that the proposed IBSC method outperforms the traditional BSC method.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114870852","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}
Pub Date : 2023-02-23DOI: 10.1109/ECCE57851.2023.10101650
M. Hussain, N. Sharmin, Sumayea Binte Shafiul
Water cycles, climate-related hazards, and agroirrigation are strongly controlled by soil moisture (SM) content. For water resource management, prediction is a key to mitigate and regulate expected economic losses and property damages. This paper compares two supervised machine learning (ML) techniques: support vector regression (SVR) and random forest (RF) to predict SM. In RStudio, various meteorological variables: temperature, relative humidity, wind speed, and rainfall are trained to estimate SM. For eight divisions, SM and weather variables are obtained from the National Aeronautics and Space Administration (NASA) Prediction of Worldwide Energy Resources (POWER). The experiments include daily observations for 39 (1982 to 2020) to develop SVR and RF models. To estimate SM from the predictive model, datasets from diverse regions: Rajshahi, Mymensingh, Chittagong, and Sylhet are utilized in training (60%) and Rangpur, Barisal, Khulna, and Dhaka are segregated for validation (40%) resulting in accuracy of 88 to 95.8%. This model further is applied to forecast daily SM for each city including two districts (Bogra and Jessore) and found slightly higher model performance for SVR (90.7%) than RF (90.1%) on average (Year: 2021). For agricultural, industrial and urban water supplies as well as drought, landslides, and river erosions can be mitigated by an accurate estimation of soil moisture. The investigations encourage for providing SM budget to public with supervised ML techniques mostly among data-sparse regions.
{"title":"Estimation of Soil Moisture with Meteorological Variables in Supervised Machine Learning Models","authors":"M. Hussain, N. Sharmin, Sumayea Binte Shafiul","doi":"10.1109/ECCE57851.2023.10101650","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101650","url":null,"abstract":"Water cycles, climate-related hazards, and agroirrigation are strongly controlled by soil moisture (SM) content. For water resource management, prediction is a key to mitigate and regulate expected economic losses and property damages. This paper compares two supervised machine learning (ML) techniques: support vector regression (SVR) and random forest (RF) to predict SM. In RStudio, various meteorological variables: temperature, relative humidity, wind speed, and rainfall are trained to estimate SM. For eight divisions, SM and weather variables are obtained from the National Aeronautics and Space Administration (NASA) Prediction of Worldwide Energy Resources (POWER). The experiments include daily observations for 39 (1982 to 2020) to develop SVR and RF models. To estimate SM from the predictive model, datasets from diverse regions: Rajshahi, Mymensingh, Chittagong, and Sylhet are utilized in training (60%) and Rangpur, Barisal, Khulna, and Dhaka are segregated for validation (40%) resulting in accuracy of 88 to 95.8%. This model further is applied to forecast daily SM for each city including two districts (Bogra and Jessore) and found slightly higher model performance for SVR (90.7%) than RF (90.1%) on average (Year: 2021). For agricultural, industrial and urban water supplies as well as drought, landslides, and river erosions can be mitigated by an accurate estimation of soil moisture. The investigations encourage for providing SM budget to public with supervised ML techniques mostly among data-sparse regions.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114192270","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}
Pub Date : 2023-02-23DOI: 10.1109/ECCE57851.2023.10101647
Sharith Dhar, Md. Saiful Islam
The industrial revolution has increased the use of induction motors enormously. Today's soft starter is used for controlling starting current, acceleration torque, and acceleration time of the induction motor. Intelligent techniques are used in soft starters for controlling starting parameters of the induction motor smoothly. But developed intelligent algorithm based soft starter takes more acceleration time, and due to this induction motor can not accelerate the load properly during the starting period. To solve this problem fuzzy logic-based soft starter is proposed in this paper. This proposed starting technique reaches the target through its instinctive decision making capability. The fuzzy logic controller takes stator phase current and torque from the three phase induction motor (IM) and gives firing angles to the thyristor unit in the soft starter by using the Mamdani fuzzy inference system and the mean of maximum method in defuzzification. The proposed technique accelerates the IM with the load smoothly by decreasing acceleration time. The proposed intelligent soft starter reduces the starting current of IM with a Direct on line (DOL) starting technique by more than 10% at the constant load and also provides proper acceleration torque. The proposed soft starter provides a better response compared with another method.
{"title":"Fuzzy Logic-based Soft Starter for Controlling Starting Parameters of Induction Motor","authors":"Sharith Dhar, Md. Saiful Islam","doi":"10.1109/ECCE57851.2023.10101647","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101647","url":null,"abstract":"The industrial revolution has increased the use of induction motors enormously. Today's soft starter is used for controlling starting current, acceleration torque, and acceleration time of the induction motor. Intelligent techniques are used in soft starters for controlling starting parameters of the induction motor smoothly. But developed intelligent algorithm based soft starter takes more acceleration time, and due to this induction motor can not accelerate the load properly during the starting period. To solve this problem fuzzy logic-based soft starter is proposed in this paper. This proposed starting technique reaches the target through its instinctive decision making capability. The fuzzy logic controller takes stator phase current and torque from the three phase induction motor (IM) and gives firing angles to the thyristor unit in the soft starter by using the Mamdani fuzzy inference system and the mean of maximum method in defuzzification. The proposed technique accelerates the IM with the load smoothly by decreasing acceleration time. The proposed intelligent soft starter reduces the starting current of IM with a Direct on line (DOL) starting technique by more than 10% at the constant load and also provides proper acceleration torque. The proposed soft starter provides a better response compared with another method.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"107 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122145469","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}
Pub Date : 2023-02-23DOI: 10.1109/ECCE57851.2023.10101581
M. Hasan, Md. Ali Hossain, Azmain Yakin Srizon, Abu Sayeed
Bengali, the seventh most spoken language in the world by the number of speakers, doesn't have a well-established Optical Character Recognition (OCR) system for handwritten texts. One of the major reasons behind this lacking is contributed to having no complete conjuncts database. No dataset available today covers all the conjunct characters that are used by authors around the globe. In this research, we prepared a complete dataset consisting of 292 consonant conjunct characters, which is the biggest consonant conjunct character dataset to date by the number of classes available in the literature to our knowledge. We applied Big Transfer-based M-ResNet-101x3 Deep Convolutional Neural Network (DCNN) which achieves 91.32% accuracy that outperforms the baseline EfficientNetB7 approach which obtained 81.05% accuracy.
{"title":"juktoMala: A Handwritten Bengali Consonant Conjuncts Dataset for Optical Character Recognition Using BiT-based M-ResNet-101x3 Architecture","authors":"M. Hasan, Md. Ali Hossain, Azmain Yakin Srizon, Abu Sayeed","doi":"10.1109/ECCE57851.2023.10101581","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101581","url":null,"abstract":"Bengali, the seventh most spoken language in the world by the number of speakers, doesn't have a well-established Optical Character Recognition (OCR) system for handwritten texts. One of the major reasons behind this lacking is contributed to having no complete conjuncts database. No dataset available today covers all the conjunct characters that are used by authors around the globe. In this research, we prepared a complete dataset consisting of 292 consonant conjunct characters, which is the biggest consonant conjunct character dataset to date by the number of classes available in the literature to our knowledge. We applied Big Transfer-based M-ResNet-101x3 Deep Convolutional Neural Network (DCNN) which achieves 91.32% accuracy that outperforms the baseline EfficientNetB7 approach which obtained 81.05% accuracy.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122151778","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}
Pub Date : 2023-02-23DOI: 10.1109/ECCE57851.2023.10101595
G. Hasanuzzaman, S. Iezekiel, A. Kanno
A dual-loop optoelectronic oscillator using a polymer-based modulator is demonstrated at 94.5 GHz. The measured single side band phase noise is -70 dBc/Hz at 10kHz offset frequency. A value of 40 dB is achieved for side mode suppression.
{"title":"94.5 GHz Dual-loop Optoelectronic Oscillator","authors":"G. Hasanuzzaman, S. Iezekiel, A. Kanno","doi":"10.1109/ECCE57851.2023.10101595","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101595","url":null,"abstract":"A dual-loop optoelectronic oscillator using a polymer-based modulator is demonstrated at 94.5 GHz. The measured single side band phase noise is -70 dBc/Hz at 10kHz offset frequency. A value of 40 dB is achieved for side mode suppression.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129129031","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}
Pub Date : 2023-02-23DOI: 10.1109/ECCE57851.2023.10101528
F. Chowdhury, Tania Noor, Md. Saiful Islam, Md Khorshed Alam
A brain tumor is an uncommon form of body cell proliferation. The most difficult tasks in the medical profession are to identify and categorize brain tumors. A person's life may be at risk if the brain tumor is not immediately identified or diagnosed. In this proposed method, an artificial neural network (ANN)-based technique can classify brain tumors accurately. Firstly, the images are normalized using the scaling process. Then the normalized images are segmented using the watershed algorithm. After that, the seven statistical features are extracted and then applied as input to the ANN classifier for the classification of the brain tumors. The experimental result of the proposed method provides an accuracy result of 95.8% which is better than modern state-of-the-art methods. Furthermore, compared to other contemporary techniques, the chosen seven statistical features are comparably few in illustrating this performance.
{"title":"Brain Tumor Classification Using Watershed Segmentation with ANN Classifier","authors":"F. Chowdhury, Tania Noor, Md. Saiful Islam, Md Khorshed Alam","doi":"10.1109/ECCE57851.2023.10101528","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101528","url":null,"abstract":"A brain tumor is an uncommon form of body cell proliferation. The most difficult tasks in the medical profession are to identify and categorize brain tumors. A person's life may be at risk if the brain tumor is not immediately identified or diagnosed. In this proposed method, an artificial neural network (ANN)-based technique can classify brain tumors accurately. Firstly, the images are normalized using the scaling process. Then the normalized images are segmented using the watershed algorithm. After that, the seven statistical features are extracted and then applied as input to the ANN classifier for the classification of the brain tumors. The experimental result of the proposed method provides an accuracy result of 95.8% which is better than modern state-of-the-art methods. Furthermore, compared to other contemporary techniques, the chosen seven statistical features are comparably few in illustrating this performance.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129381422","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}
Pub Date : 2023-02-23DOI: 10.1109/ECCE57851.2023.10101619
Songyang Lyu, M. Chowdhury, Ray C. C. Cheung
ElectroEncephaloGram (EEG) is associated with multiple functions, including communications with neurons, organic monitoring, and interactions with external stimuli. By decoding EEG signals, certain human activities such as sleeping, brain diseases, motor imagery, movement of limbs, and others can be observed and controlled through the brain-computer interface (BCI). Therefore, it is vital to efficiently process EEG signals with a robust and accurate system to build BCI systems with powerful applications. However, as a weak biosignal, EEG demands a fast-reaction system signal processing with high accuracy and sensitivity. In this work, a hardware/software co-design network based on Extreme Learning Machine (ELM) is introduced for the classification of certain actions, motor imagery of the human brain. This system is based on an optimized Hierarchical Extreme Learning Machine (H-ELM) on the software layer. The proposed method has advantages over previous designs with an accuracy of 90.3%. It also improves the training speed by around 25X compared to conventional methods. The software model is also translated into efficient FPGA hardware design to maintain high computation efficiency and reduce power consumption for biomedical applications.
{"title":"Efficient Hardware and Software Co-design for EEG Signal Classification based on Extreme Learning Machine","authors":"Songyang Lyu, M. Chowdhury, Ray C. C. Cheung","doi":"10.1109/ECCE57851.2023.10101619","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101619","url":null,"abstract":"ElectroEncephaloGram (EEG) is associated with multiple functions, including communications with neurons, organic monitoring, and interactions with external stimuli. By decoding EEG signals, certain human activities such as sleeping, brain diseases, motor imagery, movement of limbs, and others can be observed and controlled through the brain-computer interface (BCI). Therefore, it is vital to efficiently process EEG signals with a robust and accurate system to build BCI systems with powerful applications. However, as a weak biosignal, EEG demands a fast-reaction system signal processing with high accuracy and sensitivity. In this work, a hardware/software co-design network based on Extreme Learning Machine (ELM) is introduced for the classification of certain actions, motor imagery of the human brain. This system is based on an optimized Hierarchical Extreme Learning Machine (H-ELM) on the software layer. The proposed method has advantages over previous designs with an accuracy of 90.3%. It also improves the training speed by around 25X compared to conventional methods. The software model is also translated into efficient FPGA hardware design to maintain high computation efficiency and reduce power consumption for biomedical applications.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127990144","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}
Pub Date : 2023-02-23DOI: 10.1109/ECCE57851.2023.10101516
Mohammad Hanif, N. Mohammad, K. Biswas, Bijoy Harun
To solve the optimization problem of Economic Load Dispatch (ELD), a number of metaheuristic approaches have already been implemented, exhibiting substantial improvement over the conventional technique. Despite this, due to the global energy crisis, research in ELD still continues to garner considerable interest. In this study, the Seagull Optimization Algorithm (SOA), a recently developed swarm intelligence technique, is applied in ELD. As the SOA algorithm has never been utilized in the ELD, it is important to investigate its efficacy and validity in this domain. Here, two case studies of ELD incorporating 6 and 10 generator units are implemented employing SOA. What's more, the performance of SOA in ELD is compared with respect to three other previously applied metaheuristics algorithms. Results indicate that SOA is a potential algorithm capable of handling the practical optimization challenge of ELD problem more effectively, especially in large power plants having more than 6 units.
{"title":"Seagull Optimization Algorithm for Solving Economic Load Dispatch Problem","authors":"Mohammad Hanif, N. Mohammad, K. Biswas, Bijoy Harun","doi":"10.1109/ECCE57851.2023.10101516","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101516","url":null,"abstract":"To solve the optimization problem of Economic Load Dispatch (ELD), a number of metaheuristic approaches have already been implemented, exhibiting substantial improvement over the conventional technique. Despite this, due to the global energy crisis, research in ELD still continues to garner considerable interest. In this study, the Seagull Optimization Algorithm (SOA), a recently developed swarm intelligence technique, is applied in ELD. As the SOA algorithm has never been utilized in the ELD, it is important to investigate its efficacy and validity in this domain. Here, two case studies of ELD incorporating 6 and 10 generator units are implemented employing SOA. What's more, the performance of SOA in ELD is compared with respect to three other previously applied metaheuristics algorithms. Results indicate that SOA is a potential algorithm capable of handling the practical optimization challenge of ELD problem more effectively, especially in large power plants having more than 6 units.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"2014 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128027491","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}
Pub Date : 2023-02-23DOI: 10.1109/ECCE57851.2023.10101584
Md. Hasanul Bari, Tanver Ahmed, M. I. Afjal, A. M. Nitu, Md. Palash Uddin, Md Abu Marjan
The remote sensing hyperspectral image (HSI) consists of hundreds of narrow and adjoining spectral bands. It carries a lot of significant information about the earth's objects. However, the use of all HSI bands leads to higher misclassification. Band reduction is a potential solution to resolve this issue, where feature selection and feature extraction methods are commonly accomplished for the reduction of bands. One of the most commonly used unsupervised feature extraction techniques is the Principal Component Analysis (PCA). But it fails to bring out the local intrinsic information from the HSI as it ponders only the global variation of the data. This problem can be addressed by the Segmented PCA (SPCA) which exploits both the global and local variance of the data by partitioning it into highly correlated blocks. Beside, another unsupervised feature extraction technique named Nonnegative Matrix Factorization (NMF) is also applied for HSI by approximating the data in a low-dimensional subspace. In this paper, we propose a feature extraction method, named Segmented Nonnegative Matrix Factorization (SNMF), performing NMF on the segmented strongly correlated blocks of HSI data. The efficacy of the proposed method is compared with PCA, NMF, and SPCA on the Indian Pines dataset with a support vector machine classifier. The experimental result shows that SNMF (89.00%) outperforms PCA (84.33%), NMF (85.37%), and SPCA (87.59%) over all classes' samples.
{"title":"Segmented Nonnegative Matrix Factorization for Hyperspectral Image Classification","authors":"Md. Hasanul Bari, Tanver Ahmed, M. I. Afjal, A. M. Nitu, Md. Palash Uddin, Md Abu Marjan","doi":"10.1109/ECCE57851.2023.10101584","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101584","url":null,"abstract":"The remote sensing hyperspectral image (HSI) consists of hundreds of narrow and adjoining spectral bands. It carries a lot of significant information about the earth's objects. However, the use of all HSI bands leads to higher misclassification. Band reduction is a potential solution to resolve this issue, where feature selection and feature extraction methods are commonly accomplished for the reduction of bands. One of the most commonly used unsupervised feature extraction techniques is the Principal Component Analysis (PCA). But it fails to bring out the local intrinsic information from the HSI as it ponders only the global variation of the data. This problem can be addressed by the Segmented PCA (SPCA) which exploits both the global and local variance of the data by partitioning it into highly correlated blocks. Beside, another unsupervised feature extraction technique named Nonnegative Matrix Factorization (NMF) is also applied for HSI by approximating the data in a low-dimensional subspace. In this paper, we propose a feature extraction method, named Segmented Nonnegative Matrix Factorization (SNMF), performing NMF on the segmented strongly correlated blocks of HSI data. The efficacy of the proposed method is compared with PCA, NMF, and SPCA on the Indian Pines dataset with a support vector machine classifier. The experimental result shows that SNMF (89.00%) outperforms PCA (84.33%), NMF (85.37%), and SPCA (87.59%) over all classes' samples.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129425128","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 spread of false health news and its dis-semination on the internet has become a major concern, due to its potential to have disastrous effects. To detect it, numerous approaches have been attempted. However, we are aware of very few studies that have sought to identify health related false information in Bangla. In this study, we have analyzed the performance of various Machine Learning and Deep Learning approaches in detecting Bangla health-related misinformation that is available online. We have created a comprehensive data repository, consisting more than 5000 data, manually annotated to two fixed categories. Several supervised machine learning classifiers and Deep Learning algorithms have been employed in this experiment to detect fake health news at the textual level. Our experiment achieves maximum accuracy of 88% in the Passive Aggressive approach and 89% in the Bi-LSTM approach. We believe that our dataset is a significant collection of health-related data in Bangla. It may open up new perspectives for the analysis of Bangla-language and health misinformation detection.
{"title":"A Machine Learning and Deep Learning Based Approach to Detect Inaccurate Health Information in Bengali Language","authors":"Nusrat Taki, Eshatur Showan, Umratul Chowdhury, Farzana Tasnim","doi":"10.1109/ECCE57851.2023.10101612","DOIUrl":"https://doi.org/10.1109/ECCE57851.2023.10101612","url":null,"abstract":"The spread of false health news and its dis-semination on the internet has become a major concern, due to its potential to have disastrous effects. To detect it, numerous approaches have been attempted. However, we are aware of very few studies that have sought to identify health related false information in Bangla. In this study, we have analyzed the performance of various Machine Learning and Deep Learning approaches in detecting Bangla health-related misinformation that is available online. We have created a comprehensive data repository, consisting more than 5000 data, manually annotated to two fixed categories. Several supervised machine learning classifiers and Deep Learning algorithms have been employed in this experiment to detect fake health news at the textual level. Our experiment achieves maximum accuracy of 88% in the Passive Aggressive approach and 89% in the Bi-LSTM approach. We believe that our dataset is a significant collection of health-related data in Bangla. It may open up new perspectives for the analysis of Bangla-language and health misinformation detection.","PeriodicalId":131537,"journal":{"name":"2023 International Conference on Electrical, Computer and Communication Engineering (ECCE)","volume":"263 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133579897","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：