Research Papers
In this paper, we propose different deep convolutional neural network(CNN) architectures to extract features from images of chest X-rays and classify the images to detect the presence of pneumonia in a person with higher accuracy. For this research work, we have used the real-world dataset contributed by the national library of Medicine for research work.
The objective of the research is to create a novel approach for summarizing multiple documents on a particular topic, aiming to save time and effort in retrieving information. This method combines extractive summarization, which identifies important sentences or phrases to capture the main ideas, with sentiment analysis to detect contrasting opinions in the documents. By gathering document URLs and generating individual summaries using extractive techniques, a comprehensive summary is generated that encompasses diverse perspectives. The effectiveness of this proposed method is compared to existing approaches to ensure it delivers accurate and informative summaries, simplifying the search for relevant information while maintaining the fidelity and significance of the original content.
The research aims to develop a new method for summarizing multiple documents on a specific topic to enhance search efficiency. It combines extractive summarization, which selects key sentences or phrases to create a concise summary, with sentiment analysis to identify differing opinions among the documents. The method retrieves document URLs, generates individual summaries using extractive techniques, and then creates a comprehensive summary that accounts for differing viewpoints. The proposed method's performance will be compared to existing methods to assess its effectiveness. Overall, this research offers a promising approach to generating accurate and relevant summaries that save time and effort in information retrieval.
This study examined social media data from Twitter, news articles, and Reddit to investigate the relationship between the COVID-19 pandemic and the increase in domestic abuse. The analysis indicates that domestic violence acts as an opportunistic infection, thriving in the current pandemic conditions. Evaluating tweet sentiments regarding domestic violence across various social media platforms is a significant concern. The study employed topic modeling techniques such as Latent Semantic Analysis (LSA), Hierarchical Dirichlet Process (HDP), and Latent Dirichlet Allocation (LDA) to gain deeper insights into the surge of domestic violence on social media. The aim is to propose a comprehensive approach to address this issue effectively.
​This Research paper attempts to compare the performance of ANNs, with more contemporary models like, Random Forest, Logistic Regression, Naïve Bayes, KNN and Support Vector Machine on the Thyroid Disease Dataset.
The research has been implemented to classify the ground level-based ozone data based on complex machine learning models to solve this problem. The data has two classes as the target; here, the ozone day is represented by class one, and the non-ozone day is defined as class zero. Ground-level ozone is a hazardous pollutant that silently kills humans, animals, and plants.
