Projects
1. Legal Intelligence
Organisation: UNT- Information Science Lab - Discovery Park
Mentor: Dr. Junhua Ding
Project: To build a Legal Intelligence system for lawyers in identifying and judging the legal cases In a much fast and easy manner from the free online resources like Courtlistner(A huge collection of legal documents website) and fetching the data from the websites using Python Web scrapping technique and Using the Machine learning and Google latest search algorithm BERT (Bidirectional Encoder Representations from Transformers) filtering the text and training the model using sample data available and predicting the any new legal cases should fall under any of the five labels (Rule, Law, Fact, Analysis, Conclusion).
2.Characterizing the Psychological Resilience and Behaviors Facing Natural Disasters
Organisation: UNT- Information Science Lab - Discovery Park
Mentor: Dr. Lingzi Hong
Project: In this project, I used Twitter data to model and analyze communication behaviors using Google Translate API for converting the strings from different languages to English and Python Regular Expressions. Previous work has used crisis taxonomies or manually labeling methods to understand the content. However, such methods usually require extra efforts to find insights related to specific events. We are using a semi-automatic framework which is composed of two major parts to go beyond the limitations of previous methods: a) Natural Language Processing , a language model that automatically evaluate the relevance of tweets to disasters. b) topics models that learn the communication topics of citizens and local governments. With these identified topics, We analyzed what people tweeted at various spatiotemporal scales, therefore to reveal the communication gaps between citizens and local governments. The result helps to understand the urban-rural divide in response, and the communication difference due to the severity of the disaster.
3. Candidates Rank prediction using Python Deep Learning for his contribution towards Coding Community.
As the current interview scenario is biased towards “candidate’s performance during the 3-hour interview” and doesn’t take other factors into account, such as the candidate’s competitive coding abilities, contribution towards the developer community, and so on, so, In this project, I used Python and Deep Learning algorithms which helps in scraping, aggregating, and assessing the technical ability of a candidate based on publicly available sources like Github, StackOverflow, CodeChef, Codebuddy, Codeforces, Hackerearth, SPOJ, GitAwards, etc. Once the data is collected, the algorithm then defines a comprehensive scoring system that grades the candidate’s technical capabilities based on the following factors: Ranking Number of Problems Solved, Activity, Reputation, Contribution, Followers. The candidate is then assigned a scored out of 100. This helps the interviewer get a full view of a candidate’s abilities and hence make an unbiased, informed decision.
4. Insurance Fraud Claims Detection using Machine Learning
In this project , In order to overcome the traditional approach for fraud detection based on developing heuristics around fraud indicator for insurance, I Developed a Machine learning model using Decision tree(DT), Random forest(RF), Naive Bayes(NB) algorithms by training the algorithm with train and test datasets to detect the fraud claims in the auto/vehicle insurance and compared the performance by calculating the confusion matrix, which help to calculate accuracy, precision, and recall and achieved the accuracy of about 85% in predicting the fraud claims.
5. Erie Data Hub
Environment: MarkLogic 8.0, 9.0, Windows
Client: Erie Insurance Group
Project Description: Erie Data Hub is the new data platform which is being built for different portfolios in Erie Insurance Group like Claims, Billing, Agency, Customer and ESignature. MarkLogic is used as the Database for Creating the Data Hub using MarkLogic Data Hub Framework. Here MarkLogic is used as Operational Data Store.
6. EDW Release 1
Environment: Hortonworks, Windows
Client: Pekin Insurance
Project Description: EDW Release 1 in Pekin Insurance project is the data platform which is used to migrate the data from mainframe to Hortonworks.
7. LogFile Analysis Using Hadoop
Environment: Ubuntu 16.04, Hadoop-2.7.3, Hive-1.2.1, Hbase-1.2.4, Pig- 0.16.0, NetBeans IDE.
Project Description: Analyzing log data and acquire knowledge from it for managing the large chunk of web log data using Hadoop MapReduce and to construct log analysis system which depicts trends based on the users browsing mode which facilitates handling of heterogeneous query execution on a log file.
8.KID SAFEGUARD
Environment: Arduino Uno R3-ATmega328P, Eclipse IDE.
Project Description: It is an IOT plus Android project where the safety of the child is provided by a wearable band to protect from any restricted areas in the home with the help of XBee transmitter and receiver using the buzzer to alert and a mobile application to send an instant notification to the parent/guardian. Our problem deals with protecting kids of age 6 months-3years and mentally retarded people in the home from danger places such as swimming pools, wells, cooking places, failed sockets,stairs by which they may die or injured.