Claim Genius, Iselin, New Jersey June – August 2019 Data Science Intern Task: Create a prototype that predicts the parts of the car that are affected based on the severity of an accident. • Developed a machine learning application that uses Market Basket Analysis and Association Rule Mining on car claim data. • Data Cleaning and Analysis – To remove unwanted symbols, words, etc from the data to process it. • Finding Data Anomalies and verification – To check if the data is correctly labeled or not given by clients. • Rule Generation – Creating rules with which we can figure out other parts of the car that are most likely to be damaged due to the accident. • Refining Rules – Limiting the number and size of the rules and, finding relevant rules. • Data Visualization and Analysis – Interactive visualization of the many association rules based on different criteria.
Projects
Brain Tumor Segmentation (CV, Python, CNN, Keras, Tensorflow) • Designed a classification cum segmentation model to detect and segment brain tumors from MRI scans. The classification was done using the Convolution Neural Network and used VGG19 to train the model for this binary problem. Created an image segmentation model further to segment the tumor in the image using Mask R-CNN.