Bulletin of Electrical Engineering and Informatics, Apr 1, 2024
Recent research has focused on opinion mining from public sentiments using natural language proce... more Recent research has focused on opinion mining from public sentiments using natural language processing (NLP) and machine learning (ML) techniques. Transformer-based models, such as bidirectional encoder representations from transformers (BERT), excel in extracting semantic information but are resourceintensive. Google's new research, mixing tokens with fourier transform, also known as FNet, replaced BERT's attention mechanism with a non-parameterized fourier transform, aiming to reduce training time without compromising performance. This study fine-tuned the FNet model with a publicly available Kaggle hotel review dataset and investigated the performance of this dataset in both FNet and BERT architectures along with conventional machine learning models such as long short-term memory (LSTM) and support vector machine (SVM). Results revealed that FNet significantly reduces the training time by almost 20% and memory utilization by nearly 60% compared to BERT. The highest test accuracy observed in this experiment by FNet was 80.27% which is nearly 97.85% of BERT's performance with identical parameters. This is an open access article under the CC BY-SA license.
Security is regarded as a key concern of home automation systems since it ensures the resident... more Security is regarded as a key concern of home automation systems since it ensures the resident's comfort and well-being. As crime is piling up rapidly in today’s culture, home defense devices are very important. With the advent of innovative technology in recent years, homeowners can go outside the home with relief as modern home security systems offer ample protection againstburglars, fire, smoke, etc. This article implements a smart protection box for legitimate personnel that can help to notify when an unapproved user tries to infringe the security of necessary materials inside the box for taking immediate action. Two other main aspects of this smart security box, along with the upgraded security features are affordability and accessibility. This low-priced, low-power enabled the smart security box based on empirical research, is found to be durable and reliable in detecting and providing legal customers with security violation notifications. This forthcoming smart security b...
2021 International Conference on Information and Communication Technology for Sustainable Development (ICICT4SD)
In many developing and poor countries, people often migrate to suitable countries to earn their l... more In many developing and poor countries, people often migrate to suitable countries to earn their livelihood and support their families. Due to the ongoing pecuniary disaster that ensued because of COVID-19, many immigrants are coming back to their ancestry from different migrant-friendly countries for several reasons. In this paper, a novel approach has been proposed to segregate these countries into five vulnerability labels based on probabilistic likelihood score (LHS) and unsupervised clustering algorithms (CA). A survey dataset of returnee people including various information has been collected and leveraged as attributes in this study. Depending on the dissemination of attributes, LHS has been generated using Bayes' Theorem for each vulnerable country and three unsupervised mining algorithms (KMeans++, Agglomerative and BIRCH) have been applied to the LHS for categorization. Output labels obtained from CA are then evaluated appropriating the average LHS. Multiple performance measurement metrics (Adjusted Rand Index, Mutual Information based Score) have been consolidated to get an incisive comparison of vulnerability labels resulting from CA and expected LHS. The highest value of 0.74 has been attained as Normalized Mutual Information based Score for BIRCH clustering accompanying ample results for the remaining algorithms. The result has shown that the combined application of probabilistic LHS and unsupervised CA can be a reliable method to identify the vulnerability of different countries generally chosen by migrant people.
Research on cyberbullying detection is gaining increasing attention in recent years as both indiv... more Research on cyberbullying detection is gaining increasing attention in recent years as both individual victims and societies are greatly affected by it. Moreover, ease of access to social media platforms such as Facebook, Instagram, Twitter, etc. has led to an exponential increase in the mistreatment of people in the form of hateful messages, bullying, sexism, racism, aggressive content, harassment, toxic comment etc. Thus there is an extensive need to identify, control and reduce the bullying contents spread over social media sites, which has motivated us to conduct this research to automate the detection process of offensive language or cyberbullying. Our main aim is to build single and double ensemble-based voting model to classify the contents into two groups: ‘offensive’ or ‘non-offensive’. For this purpose, we have chosen four machine learning classifiers and three ensemble models with two different feature extraction techniques combined with various n-gram analysis on a datas...
Selection for salinity tolerance genotypes of rice based on phenotypic performance alone is less ... more Selection for salinity tolerance genotypes of rice based on phenotypic performance alone is less reliable and will delay progress in breeding. Recent advent of molecular markers, microsatellites or simple sequence repeats (SSRs) are used to find out salt tolerant rice genotypes. Three selected SSR markers; RM7075, RM336 and RM253 were used to evaluate rice genotypes for salt tolerance. Phenotypic and
Fake or misleading information detection is attracting researchers from all around the world in r... more Fake or misleading information detection is attracting researchers from all around the world in recent years as both society and the political world are greatly influenced by it. Moreover, various popular social media sites such as Twitter, Facebook, Instagram, etc. have accelerated the increase in the dissemination of rumors, false cures, conspiracy theories in the forms of posts, articles, videos, URLs during the COVID-19 pandemic. Thus there is an extensive need to find new techniques to verify or check the reliability of the online contents, which has inspired us to conduct this research to automatically detect misleading information. Our main aim is to create a three-level voting model to categorize the information into two classes: ‘real’ or ‘misleading’. Five conventional mining algorithms and three ensemble models have been deployed with two distinct feature extraction techniques accompanying multiple sets of n-gram profiles on a benchmark dataset. Our research outcome shows...
International Journal of Advanced Scientific Innovation, 2021
Security is regarded as a key concern of home automation systems since it ensures the resident's ... more Security is regarded as a key concern of home automation systems since it ensures the resident's comfort and well-being. As crime is piling up rapidly in today's culture, home defense devices are very important. With the advent of innovative technology in recent years, homeowners can go outside the home with relief as modern home security systems offer ample protection against burglars, fire, smoke, etc. This article implements a smart protection box for legitimate personnel that can help to notify when an unapproved user tries to infringe the security of necessary materials inside the box for taking immediate actions. Two other main aspects of this smart security box, along with the upgraded security features are affordability and accessibility. This low-priced, low-power enabled the smart security box based on empirical research, is found to be durable and reliable in detecting and providing legal customers with security violation notifications. This forthcoming smart security box can detect security breaches with the highest accuracy of 100% if proper light facilities can be arranged along with the warning SMS transmission within 800 milliseconds. This low-cost device has been built for only 62.5$ and can be commercially exploited for smart security technologies in the local and foreign markets.
Bulletin of Electrical Engineering and Informatics, Apr 1, 2024
Recent research has focused on opinion mining from public sentiments using natural language proce... more Recent research has focused on opinion mining from public sentiments using natural language processing (NLP) and machine learning (ML) techniques. Transformer-based models, such as bidirectional encoder representations from transformers (BERT), excel in extracting semantic information but are resourceintensive. Google's new research, mixing tokens with fourier transform, also known as FNet, replaced BERT's attention mechanism with a non-parameterized fourier transform, aiming to reduce training time without compromising performance. This study fine-tuned the FNet model with a publicly available Kaggle hotel review dataset and investigated the performance of this dataset in both FNet and BERT architectures along with conventional machine learning models such as long short-term memory (LSTM) and support vector machine (SVM). Results revealed that FNet significantly reduces the training time by almost 20% and memory utilization by nearly 60% compared to BERT. The highest test accuracy observed in this experiment by FNet was 80.27% which is nearly 97.85% of BERT's performance with identical parameters. This is an open access article under the CC BY-SA license.
Security is regarded as a key concern of home automation systems since it ensures the resident... more Security is regarded as a key concern of home automation systems since it ensures the resident's comfort and well-being. As crime is piling up rapidly in today’s culture, home defense devices are very important. With the advent of innovative technology in recent years, homeowners can go outside the home with relief as modern home security systems offer ample protection againstburglars, fire, smoke, etc. This article implements a smart protection box for legitimate personnel that can help to notify when an unapproved user tries to infringe the security of necessary materials inside the box for taking immediate action. Two other main aspects of this smart security box, along with the upgraded security features are affordability and accessibility. This low-priced, low-power enabled the smart security box based on empirical research, is found to be durable and reliable in detecting and providing legal customers with security violation notifications. This forthcoming smart security b...
2021 International Conference on Information and Communication Technology for Sustainable Development (ICICT4SD)
In many developing and poor countries, people often migrate to suitable countries to earn their l... more In many developing and poor countries, people often migrate to suitable countries to earn their livelihood and support their families. Due to the ongoing pecuniary disaster that ensued because of COVID-19, many immigrants are coming back to their ancestry from different migrant-friendly countries for several reasons. In this paper, a novel approach has been proposed to segregate these countries into five vulnerability labels based on probabilistic likelihood score (LHS) and unsupervised clustering algorithms (CA). A survey dataset of returnee people including various information has been collected and leveraged as attributes in this study. Depending on the dissemination of attributes, LHS has been generated using Bayes' Theorem for each vulnerable country and three unsupervised mining algorithms (KMeans++, Agglomerative and BIRCH) have been applied to the LHS for categorization. Output labels obtained from CA are then evaluated appropriating the average LHS. Multiple performance measurement metrics (Adjusted Rand Index, Mutual Information based Score) have been consolidated to get an incisive comparison of vulnerability labels resulting from CA and expected LHS. The highest value of 0.74 has been attained as Normalized Mutual Information based Score for BIRCH clustering accompanying ample results for the remaining algorithms. The result has shown that the combined application of probabilistic LHS and unsupervised CA can be a reliable method to identify the vulnerability of different countries generally chosen by migrant people.
Research on cyberbullying detection is gaining increasing attention in recent years as both indiv... more Research on cyberbullying detection is gaining increasing attention in recent years as both individual victims and societies are greatly affected by it. Moreover, ease of access to social media platforms such as Facebook, Instagram, Twitter, etc. has led to an exponential increase in the mistreatment of people in the form of hateful messages, bullying, sexism, racism, aggressive content, harassment, toxic comment etc. Thus there is an extensive need to identify, control and reduce the bullying contents spread over social media sites, which has motivated us to conduct this research to automate the detection process of offensive language or cyberbullying. Our main aim is to build single and double ensemble-based voting model to classify the contents into two groups: ‘offensive’ or ‘non-offensive’. For this purpose, we have chosen four machine learning classifiers and three ensemble models with two different feature extraction techniques combined with various n-gram analysis on a datas...
Selection for salinity tolerance genotypes of rice based on phenotypic performance alone is less ... more Selection for salinity tolerance genotypes of rice based on phenotypic performance alone is less reliable and will delay progress in breeding. Recent advent of molecular markers, microsatellites or simple sequence repeats (SSRs) are used to find out salt tolerant rice genotypes. Three selected SSR markers; RM7075, RM336 and RM253 were used to evaluate rice genotypes for salt tolerance. Phenotypic and
Fake or misleading information detection is attracting researchers from all around the world in r... more Fake or misleading information detection is attracting researchers from all around the world in recent years as both society and the political world are greatly influenced by it. Moreover, various popular social media sites such as Twitter, Facebook, Instagram, etc. have accelerated the increase in the dissemination of rumors, false cures, conspiracy theories in the forms of posts, articles, videos, URLs during the COVID-19 pandemic. Thus there is an extensive need to find new techniques to verify or check the reliability of the online contents, which has inspired us to conduct this research to automatically detect misleading information. Our main aim is to create a three-level voting model to categorize the information into two classes: ‘real’ or ‘misleading’. Five conventional mining algorithms and three ensemble models have been deployed with two distinct feature extraction techniques accompanying multiple sets of n-gram profiles on a benchmark dataset. Our research outcome shows...
International Journal of Advanced Scientific Innovation, 2021
Security is regarded as a key concern of home automation systems since it ensures the resident's ... more Security is regarded as a key concern of home automation systems since it ensures the resident's comfort and well-being. As crime is piling up rapidly in today's culture, home defense devices are very important. With the advent of innovative technology in recent years, homeowners can go outside the home with relief as modern home security systems offer ample protection against burglars, fire, smoke, etc. This article implements a smart protection box for legitimate personnel that can help to notify when an unapproved user tries to infringe the security of necessary materials inside the box for taking immediate actions. Two other main aspects of this smart security box, along with the upgraded security features are affordability and accessibility. This low-priced, low-power enabled the smart security box based on empirical research, is found to be durable and reliable in detecting and providing legal customers with security violation notifications. This forthcoming smart security box can detect security breaches with the highest accuracy of 100% if proper light facilities can be arranged along with the warning SMS transmission within 800 milliseconds. This low-cost device has been built for only 62.5$ and can be commercially exploited for smart security technologies in the local and foreign markets.
Uploads
Papers by Shovan Bhowmik