LATEST PROJECTS
Project | 01
Object Detection on Live Video Stream
Detects any object from the live video feed or a prerecorded video and then classifies it. Trained a neural network and then used it to classify the image(frame) coming from the live video feed. Model built is robust and is unsusceptible to the noise in the video.
Project | 02
Titanic Dataset Analysis
Analyzed the Titanic dataset to predict the number of people survived from that incident. Used statistics and probability to extract the useful information out of it and used it to fill the missing values. Used Random Forest technique to build the model.
Project | 03
Loan Eligibility Prediction
Built a model to predict weather a person is eligible for getting a loan or not. Used 14 features like his income, credit history, dependents etc. Cleaned the data, filled the missing values and extracted new meaningful features to learn a machine learning model with high accuracy.
Project | 04
Brooklyn Housing Sales Prediction
The model predicts the sales of houses in the Brooklyn area at a particular place and time. Considered various real-time factor like economy of the city, hiring rate of corporate industry, community services etc. Analyzed the relation between different factors by graphical interpretation in Tableau.
Academic Projects
Image Recognition (Python- tensor-flow)
-
Constructed a convolutional neural network to classify 50 different categories with few training data
-
Augmented new data and used dropouts to protect the model from overfitting
-
Reduced the dimensionality of input pixel data to accelerate the training of the model
Analysis of weather data (SQL & PySpark)
-
Retrieved 20 GB of weather data from Dark Sky API and stored it in Hadoop Distributed File System
-
Analyzed change in trends of temperature in different areas with time using SQL queries
Messenger (Java)
-
Developed a social networking system involving client- server communication using TCP/IP
-
Implemented sockets to establish the connection and made a graphical user-interface using java
Email Spam Filtering (Matlab)
-
Understood the content of the email using Natural Language Processing and extracted the key information
-
Developed sophisticated features by stemming; trained a model to detect spam emails using SVM
BigMart Sales Prediction
-
Built predictive model and find out the sales of each product at a particular store of BigMart
-
Cleaned the data and filled the missing values
-
Extracted new meaningful features from the given data which helped to improve the performance by 2%