Ishu Jaswal

Takes a text file as input and outputs the set of lexical chains in the text. Using the lexical chains it automatically creates a summary of the input article.

Built a text classifier to determine whether a movie review is expressing positive or negative sentiment. The data came from the website
IMDB.com. After that implemented a content based recommendation-algorithm using list of genres for a movie as the content.

Document clustering or text clustering is an application of cluster analysis to textual documents. Using this clustering mechanism and its different implementations we will focus on modelling topics and clustering the documents based on these topics. The main aim of this project is to provide an overview of some widely-used document clustering techniques. We will compare three different techniques viz. Latent Dirichlet Allocation (LDA), Latent Semantic Analysis (LSA) and Word2Vec and, analyze… Show more

Document clustering or text clustering is an application of cluster analysis to textual documents. Using this clustering mechanism and its different implementations we will focus on modelling topics and clustering the documents based on these topics. The main aim of this project is to provide an overview of some widely-used document clustering techniques. We will compare three different techniques viz. Latent Dirichlet Allocation (LDA), Latent Semantic Analysis (LSA) and Word2Vec and, analyze our results to learn which technique is better. For better analysis, we will be working over 2 datasets: Enron Email Dataset and Yelp Dataset, to carry out experiment on multiple different topics.
Show less

Using Python fetched real time data for top 10 most popular users from Github using GitHub API to perform community detection and create recommendation system using Python. The aim is to recommend new repositories to a user that he/she can mark as star per their interest. Four scripts were created to implement this: Data collection, Clustering data, Recommendation system, and Summary.

Used Twitter API to construct a social network for 4 U.S. presidential candidates using their account information. Fetched interesting information such as their friends/followers/following to construct a better network using own analysis. Used libraries such as Networkx, Matplotlib to plot the links between them and print some stats of resulting graph.

As the core of solution implementation, this project utilized multiple facets of advanced analytics (i.e., predictive modeling, statistical algorithms, data-driven segmentation, real-time scoring, etc.) to dramatically change the way our customer conducted their sales and marketing efforts. I was an integral part of – both in solution development as well as application maintenance), Software Group (Analytics product support for IBM SPSS suite of products) and IBM Cloud Services (Infrastructure… Show more

As the core of solution implementation, this project utilized multiple facets of advanced analytics (i.e., predictive modeling, statistical algorithms, data-driven segmentation, real-time scoring, etc.) to dramatically change the way our customer conducted their sales and marketing efforts. I was an integral part of – both in solution development as well as application maintenance), Software Group (Analytics product support for IBM SPSS suite of products) and IBM Cloud Services (Infrastructure provision and management of the platform, on which the entire solution implementation was hosted). Managed enormous data bases along with managing incremental periodical data load on top of history. I was involved in the analysis of the source system data, Designing and restructuring of the systems and jobs to incorporate the new data and development of the new code.
Show less