"An efficient Artificial Neural Network for Coronary Heart Disease Prediction", Volume 9, Issue XII, International Journal for Research in Applied Science and Engineering Technology (IJRASET) Page No: 1474-1483, ISSN: 2321-9653 (Impact Factor: 7.429)
DOI:https://doi.org/10.22214/ijraset.2021.39559
“Insurance Management with Premium Prediction ", Volume 9, Issue XII, International Journal for Research in Applied Science and Engineering Technology (IJRASET) Page No: 1222-1238, ISSN: 2321-9653 (Impact Factor: 7.429)
DOI:https://doi.org/10.22214/ijraset.2021.39416
PROJECTS
Topic Modeling on Customer Reviews - Yelp.com
- This Yelp.com review analysis project, driven by Python and tools like LDA, Scikit-learn, and Power BI, aimed to enhance user experience and support businesses. It involved data collection (web scraping or Yelp’s API) using Selenium and text preprocessing (spaCy, NLTK, Gensim).
- An ensemble machine learning model classified reviews, while Latent Dirichlet Allocation (LDA) uncovered topics, visualized with Matplotlib and Power BI. The results provided actionable recommendations for improving customer satisfaction, loyalty, and business performance.
Epicor-Driven Data Enlightenment: Transforming Business Intelligence
- Implemented a data-driven dashboard to enhance operational efficiency and decision-making. Integrated data from diverse sources and conducted analyses to extract insights, refining data manipulation skills. Utilized Amazon Redshift for data pipelines and Excel Macros for streamlined processing.
- Developed interactive dashboards using Power BI, Tableau & Quicksight, and in ERP systems like Epicor and SAP. Provided stakeholders with real-time visibility into key indicators, facilitating informed decisions. Wrote Business Activity Queries (BAQ) in Epicor and generated MRP dashboards, achieving a 40% increase in efficiency.
HPC Analytics Dashboard Application Development
- Engineered an application for High Performance Computing (HPC) with a theoretical peak performance of 1 Petaflop (PF) and an infrastructure comprising over 10,000 CPU cores and 44 GPU cards.
- Integrated Oracle RDMS for data storage and management. Leveraged Dash, Matplotlib, and Seaborn Python libraries and ReactJS for data visualization.
- This project entailed implementing Software Development Life Cycle methods such as Agile and Scrum , utilizing Jira for project management. Harnessing problem-solving, communication abilities, and collaborative teamwork, the analysis delivered actionable business insights.
Web Determining the Causal Inference of a new pricing strategy on customer retention rates for an online subscription service (Netflix):
- Conducted Predictive Analytics to predict customer churn and identify potential factors affecting customer retention.
- The project involved data collection, preprocessing, and performing A/B testing, followed by statistical analysis using Stata to interpret the results and determine the magnitude of the effect of the pricing strategy on customer retention rates.
Exploratory Data Analysis for Bureau of Transportation Statistics Flight Performance:
- Implemented a data pipeline, Developed a storage model in NoSQL server, Executed an algorithm using a parallel programming framework using Hadoop
- Proposed a cleaning improvement solution, Explored a big data cloud platform environment and finally created an reliable data management plan. K-Means Clustering Algorithm was implemented.
Claim Severity Prediction using Computer Vision and Machine Learning:
- Designed and implemented a state-of-the-art machine learning model utilizing Convolutional Neural Networks (CNNs) and a suite of Computer Vision libraries, including OpenCV, TensorFlow, and Detectron2, Meta AI’s platform for object detection and segmentation.
- This model accurately predicted auto insurance claim severity based on images of damaged vehicles, achieving an outstanding 95% accuracy rate in distinguishing repairable from total loss cases.
Compiler Lexical and Syntax Analyzer Phase Design for String, Char & Integer Operations
- Deployment of lexical and syntax analyzer phase of the compiler for a devised set of grammar rules and conditions for variable declaration that reports error in case of syntax misalignment using C, Flex and Bison technologies
