Jugpreet Talwar - LinkedIn | LinkedIn (original) (raw)

Activity

540 followers

Experience & Education

Licenses & Certifications

Courses

CSCI 570

CSCI 561

CSCI 585

Coursera

Coursera

Coursera

CSCI 548

CSCI 572

CSCI 544

CSCI 571

Projects

Dec 2015

The website provides a view of the restaurant's menu, promotional dishes and it's leadership where a customer can go to see the details of single dish and see as well as provide a review and rate it. The website also offers a feedback section wherein the customers can submit an overall feedback about the restaurant.
Database and website hosted using json-server. NPM and Bower are used to manage packages and angular modules. Angular factory and service are used in the model to communicate…
The website provides a view of the restaurant's menu, promotional dishes and it's leadership where a customer can go to see the details of single dish and see as well as provide a review and rate it. The website also offers a feedback section wherein the customers can submit an overall feedback about the restaurant.
Database and website hosted using json-server. NPM and Bower are used to manage packages and angular modules. Angular factory and service are used in the model to communicate with the database in a RESTful manner using ngResource. Angular templates are used to construct the web pages to render views using angular UI-Router.

Nov 2015

Implemented in Python, this project aims at classifying news articles into events defined in the CAMEO event code book when the amount of available training data for the model is scarce. Using the unlabeled news articles, ran cosine similarity analysis on this data and the labeled examples of the events based on TF-IDF scores so as to build the training data set. Used Python NLTK and SKLEARN libraries to do the similarity analysis. Implemented NAIVE BAYES to train on this data and then classify…
Implemented in Python, this project aims at classifying news articles into events defined in the CAMEO event code book when the amount of available training data for the model is scarce. Using the unlabeled news articles, ran cosine similarity analysis on this data and the labeled examples of the events based on TF-IDF scores so as to build the training data set. Used Python NLTK and SKLEARN libraries to do the similarity analysis. Implemented NAIVE BAYES to train on this data and then classify news articles. Achieved an average Fscore of 0.63 on a perfect score of 1.

Sep 2015

Text classification of data sets, Emails and IMDB reviews, to classify them as SPAM/HAM for emails and POSITIVE/NEGATIVE for reviews. Implemented Naive Bayes classification technique for the classification and compared the performance against known Machine learning classification algorithm implementations, SVM and MegaM. Divided the available data into two sets(75% for training, 25% for testing and vice versa), to train the model and test it and compared the performance of the model when more…
Text classification of data sets, Emails and IMDB reviews, to classify them as SPAM/HAM for emails and POSITIVE/NEGATIVE for reviews. Implemented Naive Bayes classification technique for the classification and compared the performance against known Machine learning classification algorithm implementations, SVM and MegaM. Divided the available data into two sets(75% for training, 25% for testing and vice versa), to train the model and test it and compared the performance of the model when more data was available to train as opposed to when less data was available. Implemented using Python.
https://github.com/JugpreetS/spam-sentiment-analysis

Mar 2015

Indexed the crawled data using Apache Solr. Used CLAVIN to augment the documents' metadata with latitude and longitude. Used MALLET and its LDA functionality for topic modeling on the Solr index. Developed and compared two sets of ranking and retrieval approaches: content-based and link based on the textual content of the documents to perform relevancy analysis.
Answered relevant scientific questions related to Geospatial properties over time in the Arctic and Antarctic. Created a Nutch Rest…
Indexed the crawled data using Apache Solr. Used CLAVIN to augment the documents' metadata with latitude and longitude. Used MALLET and its LDA functionality for topic modeling on the Solr index. Developed and compared two sets of ranking and retrieval approaches: content-based and link based on the textual content of the documents to perform relevancy analysis.
Answered relevant scientific questions related to Geospatial properties over time in the Arctic and Antarctic. Created a Nutch Rest API endpoint to generate D3-based visualization of link-based relevancy.
Other creators

Feb 2015

Crawled NASA AMD and NASA ACADIS websites using Nutch and wrote a program to determine MIME Types of the crawled content. Used Tika for Content detection to verify the identified content. Devised deduplication algorithms for exact duplicates and near duplicates from the fetched and parsed content using Simhash technique by implementing SHA-2.
Other creators

Nov 2014

An android application and equivalent responsive website used to search for real estate. A PHP application hosted at EC2 using AWS Elastic Beanstalk, calls Zillow API to fetch the property information and converts it from XML to JSON. The website parses the JSON object and renders the results using Bootstrap and JQuery. The website has integration to post the property information on the user's Facebook wall.
Technologies: Android, HTML5, CSS3, Bootstrap, JavaScript, PHP, AJAX, JSON…
An android application and equivalent responsive website used to search for real estate. A PHP application hosted at EC2 using AWS Elastic Beanstalk, calls Zillow API to fetch the property information and converts it from XML to JSON. The website parses the JSON object and renders the results using Bootstrap and JQuery. The website has integration to post the property information on the user's Facebook wall.
Technologies: Android, HTML5, CSS3, Bootstrap, JavaScript, PHP, AJAX, JSON, SimpleXML, jQuery, Facebook and ZIllow APIs, AWS, Responsive Web Design
See project

Nov 2014

A Java program that takes a set of logical clauses and builds a knowledge base, which can then be queried to check if a given sentence can be logically inferred from the information present in the knowledge base.
Extensively used String manipulation and HashMap to build a data structure to build and retrieve information from the knowledge base.

Oct 2014

Starting with a board position, this program builds a search tree of all the possible board positions that can be played between a player and the opponent and suggests the best possible move based on heuristics. It uses Min-Max game playing algorithm to make this decision, coupled with Alpha-Beta pruning to limit the search space to relevant search branches.
Ranked among the top ten students in Reversi competition organized as part of the Artificial Intelligence course.

Sep 2014

Developed an HTML-Javascript Web Application that takes the URL of an XML document containing property information, parses the XML file as a DOM structure and extracts the details of property with their attributes and displays the information in a tabular format.

Other similar profiles

Explore top content on LinkedIn

Find curated posts and insights for relevant topics all in one place.

View top content

Add new skills with these courses