Pub Date : 2022-10-07DOI: 10.1109/GCAT55367.2022.9971838
Vineeta Gulati, Neeraj Raheja, Rajneesh Gujral
Feature Extraction (EF) is considered the effective process among all the data processing steps of the classification system. In real-life applications, the reliability of a classifier is highly affected by high-dimensional irrelevant and redundant information. Hence extraction of appropriate data plays an imperative role to reduce the dimensionality and increase the performance of the classification system. Herein paper, a hybrid Principal Independent Component Analysis (PICA) technique is presented by the combination of the two most popular Principal Component Analysis (PCA) and Singular Value Decomposition (SVD) feature extraction techniques. The authors execute the proposed PICA technique with the SGD classifier of machine learning (ML) and analyze the performance by comparing the results with existing PCA, LDA, SVD, and ICA feature extraction techniques. Furthermore, to evaluate the PICA's performance, results are compared without applying any feature extraction techniques or with existing ICA, PCA, LDA, and SVD methods. The effectiveness of the presented work is better than existing work found in the literature and is considered on an improved scale of accomplished 3.94% accuracy, 1.35% Sensitivity, 7.70% Specificity, and 5.27% precision. Moreover, decrease the 42.60% RMSE and 15% dimensionality.
{"title":"Pica-A Hybrid Feature Extraction Technique Based on Principal Component Analysis and Independent Component Analysis","authors":"Vineeta Gulati, Neeraj Raheja, Rajneesh Gujral","doi":"10.1109/GCAT55367.2022.9971838","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.9971838","url":null,"abstract":"Feature Extraction (EF) is considered the effective process among all the data processing steps of the classification system. In real-life applications, the reliability of a classifier is highly affected by high-dimensional irrelevant and redundant information. Hence extraction of appropriate data plays an imperative role to reduce the dimensionality and increase the performance of the classification system. Herein paper, a hybrid Principal Independent Component Analysis (PICA) technique is presented by the combination of the two most popular Principal Component Analysis (PCA) and Singular Value Decomposition (SVD) feature extraction techniques. The authors execute the proposed PICA technique with the SGD classifier of machine learning (ML) and analyze the performance by comparing the results with existing PCA, LDA, SVD, and ICA feature extraction techniques. Furthermore, to evaluate the PICA's performance, results are compared without applying any feature extraction techniques or with existing ICA, PCA, LDA, and SVD methods. The effectiveness of the presented work is better than existing work found in the literature and is considered on an improved scale of accomplished 3.94% accuracy, 1.35% Sensitivity, 7.70% Specificity, and 5.27% precision. Moreover, decrease the 42.60% RMSE and 15% dimensionality.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"1993 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128624107","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 : 2022-10-07DOI: 10.1109/GCAT55367.2022.9972205
Shwetambari Borade, D. Kalbande, Hemangi Jakaria, Linesh Patil
Dermatological illnesses are the most serious problem in the twenty-first century, owing to a lack of awareness and the high cost of diagnosis. To address the issue, we believe that automated methods-based applications will be extremely beneficial in the early stages of diagnosis. As a result, in this research, we describe an automated approach for identifying the kind of cosmetic skin and cosmetic skin condition from photographs. Our model is created utilizing machine learning methods. The model consists of three phases: data gathering and data augmentation, features extraction, and prediction.
{"title":"An Automated approach to detect & diagnosis the type of Cosmetic Skin & its Disease using Machine Learning","authors":"Shwetambari Borade, D. Kalbande, Hemangi Jakaria, Linesh Patil","doi":"10.1109/GCAT55367.2022.9972205","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.9972205","url":null,"abstract":"Dermatological illnesses are the most serious problem in the twenty-first century, owing to a lack of awareness and the high cost of diagnosis. To address the issue, we believe that automated methods-based applications will be extremely beneficial in the early stages of diagnosis. As a result, in this research, we describe an automated approach for identifying the kind of cosmetic skin and cosmetic skin condition from photographs. Our model is created utilizing machine learning methods. The model consists of three phases: data gathering and data augmentation, features extraction, and prediction.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129645169","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 : 2022-10-07DOI: 10.1109/GCAT55367.2022.9972047
Sheily Verma
The widespread usage of computer networks has created new dangers, making it essential that security systems be made faster and more precise. In spite of the development of new security solutions, the rapid expansion of malicious activity poses a serious danger to network security. As an initial layer of protection, firewalls and other traditional security solutions are used. The use of firewalls, on the other hand, does not guarantee complete security against invasions. Such intrusion detection systems (IDSs) are highly relied upon by network managers. With machine learning (ML), it is possible to train an intrusion detection system to distinguish between normal and anomalous traffic based on historical data. It's still simpler for cybercriminals to infiltrate networks unnoticed to steal or damage information assets because of the tremendous traffic in current network infrastructures. In a real-time processing situation, classic network intrusion detection systems fail to match expectations in terms of speed and efficiency. These constraints in mind, we set out to develop new methods for designing as well as architecting network intrusion detection systems that make use of deep learning methods. Deep learning methods are the focus of this research. An intrusion detection system that uses the 1DCNN-Bi-GRU architecture as well as a KDD99 dataset will be developed based on these models' strengths in data collecting, preprocessing, extraction and classification. Model performance can be measured against the KDD99 dataset. When compared to current algorithms, the suggested 1DCNN-Bi-GRU algorithm's accuracy was 97 percent raised by roughly 15 percent, according to the testing data.
{"title":"A Novel Approach for Designing Network Intrusion Detection Systems Based on Hybrid Deep Learning Model CB-GRU","authors":"Sheily Verma","doi":"10.1109/GCAT55367.2022.9972047","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.9972047","url":null,"abstract":"The widespread usage of computer networks has created new dangers, making it essential that security systems be made faster and more precise. In spite of the development of new security solutions, the rapid expansion of malicious activity poses a serious danger to network security. As an initial layer of protection, firewalls and other traditional security solutions are used. The use of firewalls, on the other hand, does not guarantee complete security against invasions. Such intrusion detection systems (IDSs) are highly relied upon by network managers. With machine learning (ML), it is possible to train an intrusion detection system to distinguish between normal and anomalous traffic based on historical data. It's still simpler for cybercriminals to infiltrate networks unnoticed to steal or damage information assets because of the tremendous traffic in current network infrastructures. In a real-time processing situation, classic network intrusion detection systems fail to match expectations in terms of speed and efficiency. These constraints in mind, we set out to develop new methods for designing as well as architecting network intrusion detection systems that make use of deep learning methods. Deep learning methods are the focus of this research. An intrusion detection system that uses the 1DCNN-Bi-GRU architecture as well as a KDD99 dataset will be developed based on these models' strengths in data collecting, preprocessing, extraction and classification. Model performance can be measured against the KDD99 dataset. When compared to current algorithms, the suggested 1DCNN-Bi-GRU algorithm's accuracy was 97 percent raised by roughly 15 percent, according to the testing data.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127418355","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 : 2022-10-07DOI: 10.1109/GCAT55367.2022.9972184
H. N. Rakshitha, H. M. Kalpana
Visual detections of face expression have become much difficult job and yields in reduced preciseness against the automated face expression identification making use of image processing for emotion identification and take lesser time and less efforts with increased accurate outcomes. Effective identification and classification of face-based expressions are found to be beneficial in people's behavioral monitoring system. This paper presents detailed texture bound analyzes architecture framework and codebook generation for face expression detection and classification method. The proposed system acquires Local threshold-texture Patterns (LT - TP) from input face images of unique classes. The features from LT - TP were effective and the method was able to acquire discriminant data and indicate every class containing less dimensions in it. Then, unique codebooks are generated for each class of images. Finally, Classification is done by making use of multiclass typed SVM-Support Vector Machines. Experimentations are performed on Matlab on face dataset that contains minimum two unique classes namely normal and happy faces.
{"title":"Image Processing for Facial Expression using Machine Learning Algorithm","authors":"H. N. Rakshitha, H. M. Kalpana","doi":"10.1109/GCAT55367.2022.9972184","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.9972184","url":null,"abstract":"Visual detections of face expression have become much difficult job and yields in reduced preciseness against the automated face expression identification making use of image processing for emotion identification and take lesser time and less efforts with increased accurate outcomes. Effective identification and classification of face-based expressions are found to be beneficial in people's behavioral monitoring system. This paper presents detailed texture bound analyzes architecture framework and codebook generation for face expression detection and classification method. The proposed system acquires Local threshold-texture Patterns (LT - TP) from input face images of unique classes. The features from LT - TP were effective and the method was able to acquire discriminant data and indicate every class containing less dimensions in it. Then, unique codebooks are generated for each class of images. Finally, Classification is done by making use of multiclass typed SVM-Support Vector Machines. Experimentations are performed on Matlab on face dataset that contains minimum two unique classes namely normal and happy faces.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127538909","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 : 2022-10-07DOI: 10.1109/GCAT55367.2022.9971840
S. Shruthi, K. Saraswathi
The Sparse Code Multiple Access (SCMA) is one of the code-domain non-orthogonal multiple access schemes (CD-NOMA) employed in wireless communication systems. The SCMA is a multi-dimensional codebook based spreading procedure where the incoming bits of several users are mapped directly to multi-dimensional code words that are selected from sparse codebooks. The SCMA increases the spectral efficiency by allowing a large number of users to share the time and frequency resources which improves overall system performance. This has gained attention in the field of massive-machine type communication (mMtc) in 5G to satisfy huge demand on massive connectivity and data traffic. In this paper, the design and simulation of the SCMA system for the uplink scenario is carried out for 6 and 8 UEs. The Turbo coder is used for forward error correction and the Log-message passing algorithm (Log-MPA) is used as a detection scheme. The performance analysis of SCMA system is carrier out using block error rate (BLER) for 6 and 8 UEs.
{"title":"Performance Analysis of Sparse Code Multiple Access Transceiver System for Massive Machine - Type Communication","authors":"S. Shruthi, K. Saraswathi","doi":"10.1109/GCAT55367.2022.9971840","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.9971840","url":null,"abstract":"The Sparse Code Multiple Access (SCMA) is one of the code-domain non-orthogonal multiple access schemes (CD-NOMA) employed in wireless communication systems. The SCMA is a multi-dimensional codebook based spreading procedure where the incoming bits of several users are mapped directly to multi-dimensional code words that are selected from sparse codebooks. The SCMA increases the spectral efficiency by allowing a large number of users to share the time and frequency resources which improves overall system performance. This has gained attention in the field of massive-machine type communication (mMtc) in 5G to satisfy huge demand on massive connectivity and data traffic. In this paper, the design and simulation of the SCMA system for the uplink scenario is carried out for 6 and 8 UEs. The Turbo coder is used for forward error correction and the Log-message passing algorithm (Log-MPA) is used as a detection scheme. The performance analysis of SCMA system is carrier out using block error rate (BLER) for 6 and 8 UEs.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131080710","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 : 2022-10-07DOI: 10.1109/GCAT55367.2022.9972199
R. K. Madhusudhana, G. Kiranmayi
This paper presents a system which configures a web- based access monitoring system for banks. The banking system plays a very significant role in society. There is a need for monitoring the people entering the bank as a part of the procedure for providing security. The traditional monitoring system only records the video, it does not recognize the people entering. This work implements a system for recognizing the people entering the bank using a computer vision technology. With the ongoing development of image processing algorithms, the computer vision discipline of artificial intelligence launching a new era by teaching computers to comprehend and interpret the visual environment. This project proposes an efficient way to keep track of the people entering the banks using face detection. Feature extraction of the images of the employees and the customers is done. The database of the face images of employees of the bank and customers is created and stored in the system. Viola jones algorithm which uses HAAR feature extraction and cascade classifiers is used for the face detection. The face features detected is compared with the existing database and classified as employee, customer and unknown. After the face recognition the information is displayed in a webpage with date and time of entry. This webpage can be very useful in case of Security breach in banks. The face detection algorithm is implemented using OpenCV on a Raspberry pi processor with Linux operating system.
{"title":"Design and implementation of access monitoring system for banks using OpenCV and Django","authors":"R. K. Madhusudhana, G. Kiranmayi","doi":"10.1109/GCAT55367.2022.9972199","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.9972199","url":null,"abstract":"This paper presents a system which configures a web- based access monitoring system for banks. The banking system plays a very significant role in society. There is a need for monitoring the people entering the bank as a part of the procedure for providing security. The traditional monitoring system only records the video, it does not recognize the people entering. This work implements a system for recognizing the people entering the bank using a computer vision technology. With the ongoing development of image processing algorithms, the computer vision discipline of artificial intelligence launching a new era by teaching computers to comprehend and interpret the visual environment. This project proposes an efficient way to keep track of the people entering the banks using face detection. Feature extraction of the images of the employees and the customers is done. The database of the face images of employees of the bank and customers is created and stored in the system. Viola jones algorithm which uses HAAR feature extraction and cascade classifiers is used for the face detection. The face features detected is compared with the existing database and classified as employee, customer and unknown. After the face recognition the information is displayed in a webpage with date and time of entry. This webpage can be very useful in case of Security breach in banks. The face detection algorithm is implemented using OpenCV on a Raspberry pi processor with Linux operating system.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"110 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132073016","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 : 2022-10-07DOI: 10.1109/GCAT55367.2022.9971994
Sushil Karvekar, P. Joshi
The paper presents implementation of speed control strategies for Induction motor using field oriented control and adaptive notch filters. The adaptive notch filters used for developing filed oriented control include EPLL and discrete Fourier transform. Co-simulation technique is employed by using Matlab/Simulink and PSIM softwares for implementing the speed control algorithm. The simulation results of these adaptive notch filters demonstrate the computation of torque and flux components in rotating reference frame. The practicability and reliability of these algorithms are validated by changing the reference speed and torque conditions. The simulation results demonstrate the performance of field oriented control algorithm using the discrete Fourier and EPLL for tracking the given reference speed. The results show that the modified control algorithm is efficient in reference tracking with improved computational efficiency.
{"title":"Implementation and Performance Analysis of Speed Control Strategy for Induction Motor using a Novel PLL and Discrete Fourier Transform","authors":"Sushil Karvekar, P. Joshi","doi":"10.1109/GCAT55367.2022.9971994","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.9971994","url":null,"abstract":"The paper presents implementation of speed control strategies for Induction motor using field oriented control and adaptive notch filters. The adaptive notch filters used for developing filed oriented control include EPLL and discrete Fourier transform. Co-simulation technique is employed by using Matlab/Simulink and PSIM softwares for implementing the speed control algorithm. The simulation results of these adaptive notch filters demonstrate the computation of torque and flux components in rotating reference frame. The practicability and reliability of these algorithms are validated by changing the reference speed and torque conditions. The simulation results demonstrate the performance of field oriented control algorithm using the discrete Fourier and EPLL for tracking the given reference speed. The results show that the modified control algorithm is efficient in reference tracking with improved computational efficiency.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131345686","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 : 2022-10-07DOI: 10.1109/GCAT55367.2022.10059266
Harshavardhan Yadav Gangadhara, R. J, A. R
Electric vehicles are exponentially rising in the real world, due to the rise in concern about environmental degradation. Shortly, E-mobility will overtake internal combustion engine mobility. Wireless power transfer (WPT) is a rapidly growing technology, but there is development concerning in development of technology such as high frequency signals, transmission range, size, and efficiency. WPT is a convenient topology for charging Electric Vehicles (EV). A parking lot with fewer charging stations would be inconvenient to charge vehicles. Using WPT there is no need to connect the vehicle to the charging station making it user-friendly and accessible. As the number of vehicles is being connected to the grid to charge there is a great load on the grid. In this paper, the design of a wireless battery charging station with Inductive Wireless Power Transfer (IWPT) topology using renewable energies such as solar energy farm, wind energy farm, and AC grid as sources by choosing the source energy optimally which are subjected to environmental condition for charging EVs in the parking lot with the varying number of vehicles is proposed.
{"title":"Optimisation of Source Selection and Design of High Frequency LCL Based Wireless Power Transmission Array","authors":"Harshavardhan Yadav Gangadhara, R. J, A. R","doi":"10.1109/GCAT55367.2022.10059266","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.10059266","url":null,"abstract":"Electric vehicles are exponentially rising in the real world, due to the rise in concern about environmental degradation. Shortly, E-mobility will overtake internal combustion engine mobility. Wireless power transfer (WPT) is a rapidly growing technology, but there is development concerning in development of technology such as high frequency signals, transmission range, size, and efficiency. WPT is a convenient topology for charging Electric Vehicles (EV). A parking lot with fewer charging stations would be inconvenient to charge vehicles. Using WPT there is no need to connect the vehicle to the charging station making it user-friendly and accessible. As the number of vehicles is being connected to the grid to charge there is a great load on the grid. In this paper, the design of a wireless battery charging station with Inductive Wireless Power Transfer (IWPT) topology using renewable energies such as solar energy farm, wind energy farm, and AC grid as sources by choosing the source energy optimally which are subjected to environmental condition for charging EVs in the parking lot with the varying number of vehicles is proposed.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125389569","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 : 2022-10-07DOI: 10.1109/GCAT55367.2022.9972238
R. Ramachandran, Manjusha K Mohan, Subin K Sara
Increase in the number of research documents on a daily basis, we find difficulty in identifying proper documents as per our requirements. This paper discusses an effective method in document clustering using automatic keyword extraction. Keyword is the smallest unit that can convey the meaning of an entire page, it helps a user in deciding whether or not to read or skip an article. In this work, we compare different methods of keyword extraction and choose the best method of keyword extraction based on accuracy and precision. The proposed approach takes extracted keywords as input and constructs a variety of different clusters using Euclidean distance measure to group the document together. As a result, a user can conduct a keyword search and obtain the results within seconds. The use of keyword clusters reduces noise in data and consequently enhances cluster quality.
{"title":"Document Clustering Using Keyword Extraction","authors":"R. Ramachandran, Manjusha K Mohan, Subin K Sara","doi":"10.1109/GCAT55367.2022.9972238","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.9972238","url":null,"abstract":"Increase in the number of research documents on a daily basis, we find difficulty in identifying proper documents as per our requirements. This paper discusses an effective method in document clustering using automatic keyword extraction. Keyword is the smallest unit that can convey the meaning of an entire page, it helps a user in deciding whether or not to read or skip an article. In this work, we compare different methods of keyword extraction and choose the best method of keyword extraction based on accuracy and precision. The proposed approach takes extracted keywords as input and constructs a variety of different clusters using Euclidean distance measure to group the document together. As a result, a user can conduct a keyword search and obtain the results within seconds. The use of keyword clusters reduces noise in data and consequently enhances cluster quality.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126258771","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 : 2022-10-07DOI: 10.1109/GCAT55367.2022.9971905
G. Shobana, S. Sanjay, V. Saran, G. K. Vardan
Myriad of consumer complaints has subjected to the difficulty in classifying consumer's grievances. Grievances usually comprises of lengthy texts which takes lots of manpower and time. Complaints can be filed into wrong categories. Difficulty in going through every sole grievance and directing them to relevant departments is to be dealt. To solve these issues, we have an idea of using machine learning algorithms to learn and classify the complaints into their respective categories and perform sentimental analysis on the customer complaints to obtain the priority of each complaint. Python FLASK API is used to enable application interaction. The user should enter the consumer complaint in the application, and the sentimental analysis and categorization of consumer complaints is done and the accuracy of the complaint classified is displayed.
{"title":"Consumer Grievance Handler","authors":"G. Shobana, S. Sanjay, V. Saran, G. K. Vardan","doi":"10.1109/GCAT55367.2022.9971905","DOIUrl":"https://doi.org/10.1109/GCAT55367.2022.9971905","url":null,"abstract":"Myriad of consumer complaints has subjected to the difficulty in classifying consumer's grievances. Grievances usually comprises of lengthy texts which takes lots of manpower and time. Complaints can be filed into wrong categories. Difficulty in going through every sole grievance and directing them to relevant departments is to be dealt. To solve these issues, we have an idea of using machine learning algorithms to learn and classify the complaints into their respective categories and perform sentimental analysis on the customer complaints to obtain the priority of each complaint. Python FLASK API is used to enable application interaction. The user should enter the consumer complaint in the application, and the sentimental analysis and categorization of consumer complaints is done and the accuracy of the complaint classified is displayed.","PeriodicalId":133597,"journal":{"name":"2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122566490","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：