Ex-Intern at CMU, IIT-Roorkee
Deep Learning and Computer Vision
MY projects
WeedGuard : Intelligent weed classification system using YOLO
- Conducted (EDA) on the data by creating charts using seaborn and matplotlib this reduced data errors by 12% in data,
- Developed a LSTM based sequence model with Dropout Regularization to predict the price of a stock with a loss = 0.0011.
- Analysed the variable factor that affect a stock’s price increased the overall accuracy to 84.5%
Unsupervised Learning Market Strategy and Visualisation of Portfolio Return
- Designed a K-Means Clustering based model on S&P 500 stock dataset of the past 10 years and sorted them on monthly based results and performed EDA decreasing data error by 12%
- Employs concepts of Unsupervised & visualisation accuracy of 87.4%
Image Captioning System Using Flickr Dataset
- Image Caption Generating System and VGG16 pre-trained model.
- Employs concepts of CNN and RNN & LSTM.
Online Abuse Detection Using ML
- Identifying online abuse using TfidVectorizer employs nltk and sklearn library
- Full-Stack ML projct