Project Work

I have always been fascinated by the novel and groundbreaking applications of Machine Learning, Artificial Intelligence, Natural Language Processing, and Data Science. I have opted for a diverse number of courses and projects during the term-work that have helped me immensely to enhance my understanding. Unsatisfied, I keep on exploring the field more by studying recent innovations thoroughly.

I have mostly worked on projects related to Python and Machine Learning with data visualization and analysis. A few projects that I have worked on during my undergraduate study are as follows:-

Face detection using Python and OpenCV

• ->The technique which I implemented here is a specific use case of object detection technology that deals with detecting the instances of semantic objects such as humans.
• ->OpenCV grabs each frame from the webcam and then detects faces by processing each frame. I implemented it using Haarcascades which is a machine learning-based approach.

Question Answering using Self-Attention with BERT approach

• -> The main purpose of this project was to make the existing system less tedious by reducing the amount of time required to find the answer.
• ->The whole system worked on the BERT model released by Google research. The model worked on the SQUAD dataset and is trained on it.

Document Summarization with Machine Learning

• ->It consumes a lot of time in reading the entire book which has many repetitions so it would be beneficial if there is something that can summarize the document with the relevant information and removes the repeated and unnecessary information.
• ->The system summarizes by copying and rearranging passages then generates new phrases by rephrasing or using words that are not in the original text.

Personality Prediction System

• -> Five basic characteristics of individuals are taken into consideration- openness, neuroticism, conscientiousness, agreeableness, and extraversion.
• -> Model implemented to predict test data is Logistic Regression as it is an effective model for predicting output class labels for dependent categorical data.

Speech Emotion Recognition using Deep Learning

• ->Deep learning techniques provide advantages over traditional methods as it can detect complex structures and features.
• -> Studied various approaches such as SVM, MLP, CNN, RNN and lastly implemented it using Deep neural networks (DNN) with convolutional, pooling, and fully connected layers.

Customer Segmentation using K-means Clustering

• ->Division of customers into groups based on interests and characteristics
• ->It works by finding the groups based on Euclidean Distance which is a measure of distance. Then the user selects k groups to cluster and the algorithm finds the best centroids for the groups.

Stock Market Prediction

• ->To determine future market trends of stock by evaluating past data and news headlines.
• ->Developed a model using NLP that will be able to buy and sell the stock based on the predictions without human intervention.

Company Projects- Improving Cross-Border Child Protection, Improving the Lives of Cancer Patients by Identifying Existing Non-Cancer Generic Drugs, Mentor-Mentee Recommendation System