Andavarapu Sravani
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Andavarapu Sravani has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Chaitanya Chaitanya
Centurion University,Department of Anesthisia,Vizianagaram,Andhra Pradesh,India
1 shared publication - Dinesh Valluru
MLRITM Engineering College
1 shared publication - Golagani Lavanaya Devi
Department of Computer Science and Systems Engineering, Andhra University College of Engineering, Andhra University, Visakhapatnam, India
1 shared publication - Kongara Srinivasa Rao
ICFAI Foundation for Higher Education (IFHE)
1 shared publication - N. V. S. Shankar
Department of Computer Science and Engineering, Centurion University of Technology and Management, Vizianagaram 535003, Andhra Pradesh, India
1 shared publication - P. Anthony Sunny Dayal
Centurion University of Technology and Management,Department of CSE,Vizianagaram,Andhra Pradesh,India
1 shared publication
Latest Publications
- 6 June 2025
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 37 (15-17)
DSPCNN: Deep Spiking Parallel Convolutional Neural Network for Video Recommendation System
The recommendation system infers the interest of users by collecting their behaviors and producing the recommendation list. A video recommender system aims to satisfy the requirements of users based on their interests. Many traditional video recommendation systems rely on the linear relationship between users and their neighbors, often overlooking higher‐order relationships among users, which can negatively affect recommendation accuracy. Here, a Deep Spiking Parallel Convolutional Neural Network (DSPCNN) is developed for a video recommendation system. Initially, the user‐video accumulation matrix is calculated. Thereafter, the video grouping is conducted employing the Fuzzy Local Information Cluster means (FLICM). Subsequently, bi‐level matching between the user query and the video group is performed using squared‐chord distance and Matusita distance, allowing for the retrieval of the user‐preferred video. Finally, the video recommendation is performed using the proposed DSPCNN. Moreover, the merging of Spiking Deep Residual Network (S‐ResNet) and Parallel Convolutional Neural Network (PCNN) creates the DSPCNN. Here, the video having maximum reach metrics is considered as the recommended one, which is newly developed based on click rate, number of likes, follows, comments, and forwards. Finally, the recommended video is provided to the users. It can be identified that DSPCNN has obtained an accuracy of 91%, a mean squared error (MSE) of 0.070, and a root mean squared error (RMSE) of 0.265. The source code is “https://github.com/SravaniAndavar/DSPCNN.git.”
- 6 June 2025
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 37 (15-17)
DSPCNN: Deep Spiking Parallel Convolutional Neural Network for Video Recommendation System
The recommendation system infers the interest of users by collecting their behaviors and producing the recommendation list. A video recommender system aims to satisfy the requirements of users based on their interests. Many traditional video recommendation systems rely on the linear relationship between users and their neighbors, often overlooking higher‐order relationships among users, which can negatively affect recommendation accuracy. Here, a Deep Spiking Parallel Convolutional Neural Network (DSPCNN) is developed for a video recommendation system. Initially, the user‐video accumulation matrix is calculated. Thereafter, the video grouping is conducted employing the Fuzzy Local Information Cluster means (FLICM). Subsequently, bi‐level matching between the user query and the video group is performed using squared‐chord distance and Matusita distance, allowing for the retrieval of the user‐preferred video. Finally, the video recommendation is performed using the proposed DSPCNN. Moreover, the merging of Spiking Deep Residual Network (S‐ResNet) and Parallel Convolutional Neural Network (PCNN) creates the DSPCNN. Here, the video having maximum reach metrics is considered as the recommended one, which is newly developed based on click rate, number of likes, follows, comments, and forwards. Finally, the recommended video is provided to the users. It can be identified that DSPCNN has obtained an accuracy of 91%, a mean squared error (MSE) of 0.070, and a root mean squared error (RMSE) of 0.265. The source code is “https://github.com/SravaniAndavar/DSPCNN.git.”
- 19 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Web Based Legal Text Summarizer
The legal industry is currently grappling with the issue of a growing backlog of unresolved cases. This paper aims to address the problem by proposing a web-based tool designed to generate extractive summaries from selected legal documents. These summaries can support decision-making within the judicial process. The primary goal of the tool is to reduce the time and effort legal professionals spend reviewing case files, thereby streamlining workflows. The key objective is to identify cases that can be resolved or settled more efficiently, allowing legal practitioners and courts to focus their attention on more complex matters. To achieve this, the tool must generate comprehensive summaries of key elements within the case documents. Three unsupervised and two supervised learning techniques for summarizing legal texts were investigated during the development process. Of these, the TextRank, an unsupervised learning algorithm, delivered optimal results. Consequently, this method was implemented in the tool to ensure effective and accurate summarization of lengthy legal documents. The system helps streamline the handling of pending case files and assists decision-makers in accelerating their work by summarizing important case details such as evidence and legal sections. It ensures no crucial statements are overlooked, which could be critical for the case. Additionally, a web application was developed to implement the TextRank method, automating the process. The generated summaries are stored in a database for future reference, providing a useful resource for ongoing and future legal matters.
- 19 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Web Based Legal Text Summarizer
The legal industry is currently grappling with the issue of a growing backlog of unresolved cases. This paper aims to address the problem by proposing a web-based tool designed to generate extractive summaries from selected legal documents. These summaries can support decision-making within the judicial process. The primary goal of the tool is to reduce the time and effort legal professionals spend reviewing case files, thereby streamlining workflows. The key objective is to identify cases that can be resolved or settled more efficiently, allowing legal practitioners and courts to focus their attention on more complex matters. To achieve this, the tool must generate comprehensive summaries of key elements within the case documents. Three unsupervised and two supervised learning techniques for summarizing legal texts were investigated during the development process. Of these, the TextRank, an unsupervised learning algorithm, delivered optimal results. Consequently, this method was implemented in the tool to ensure effective and accurate summarization of lengthy legal documents. The system helps streamline the handling of pending case files and assists decision-makers in accelerating their work by summarizing important case details such as evidence and legal sections. It ensures no crucial statements are overlooked, which could be critical for the case. Additionally, a web application was developed to implement the TextRank method, automating the process. The generated summaries are stored in a database for future reference, providing a useful resource for ongoing and future legal matters.
- Open Access
- 1 November 2024
- Published by Institute of Advanced Engineering and Science in Indonesian Journal of Electrical Engineering and Computer Science
- Vol. 36 (2) , 1273-1283
Phishing website detection using novel integration of BERT and XLNet with deep learning sequential models
Phishing websites pose a significant threat to online security, necessitating robust detection mechanisms to safeguard users' sensitive information. This study explores the efficacy of various deep learning architectures for phishing website detection. Initially, traditional sequential models, including recurrent neural networks (RNN), long short-term memory (LSTM), and gated recurrent unit (GRU), achieve accuracies of 95%, 96%, and 96.5%, respectively, on a curated dataset. Building upon these results, hybrid architectures that combine the strengths of traditional sequential models with state-of-the-art language representation models, bidirectional encoder representations from transformers (BERT) and XLNet, are investigated. Combinations such as RNN with BERT, BERT with LSTM, BERT with GRU, RNN with XLNet, XLNet with LSTM, and XLNet with GRU are evaluated. Through experimentation, accuracies of 94.5%, 96.5%, 96.1%, 95.7%, 97.4%, and 97%, respectively, are achieved, demonstrating the effectiveness of hybrid deep learning architectures in enhancing phishing detection performance. These findings contribute to advancing the state-of-the-art in cybersecurity practices and underscore the importance of leveraging diverse model types to combat online threats effectively.
- 1 November 2024
- Published by Institute of Advanced Engineering and Science in Indonesian Journal of Electrical Engineering and Computer Science
- Vol. 36 (2) , 1273-1283
Open AccessPhishing website detection using novel integration of BERT and XLNet with deep learning sequential models
Phishing websites pose a significant threat to online security, necessitating robust detection mechanisms to safeguard users' sensitive information. This study explores the efficacy of various deep learning architectures for phishing website detection. Initially, traditional sequential models, including recurrent neural networks (RNN), long short-term memory (LSTM), and gated recurrent unit (GRU), achieve accuracies of 95%, 96%, and 96.5%, respectively, on a curated dataset. Building upon these results, hybrid architectures that combine the strengths of traditional sequential models with state-of-the-art language representation models, bidirectional encoder representations from transformers (BERT) and XLNet, are investigated. Combinations such as RNN with BERT, BERT with LSTM, BERT with GRU, RNN with XLNet, XLNet with LSTM, and XLNet with GRU are evaluated. Through experimentation, accuracies of 94.5%, 96.5%, 96.1%, 95.7%, 97.4%, and 97%, respectively, are achieved, demonstrating the effectiveness of hybrid deep learning architectures in enhancing phishing detection performance. These findings contribute to advancing the state-of-the-art in cybersecurity practices and underscore the importance of leveraging diverse model types to combat online threats effectively.