5 assignments made as a part of the course BITS F464
1: Exploratory data analysis on a dataset.
2: CART, Random Forest, and Gradient boosting Classifiers on a dataset.
3: LDA, Linear and Logistic Regression on a dataset.
4: The goal of this assignment was to train two models—Naive Bayes and a Perceptron—to predict the type of diabetes based on input features (age, blood sugar level (Glucose), insulin level, and BMI). I compare the performance of both the models using evaluation metrics such as accuracy, precision,recall, and F1-score. Additionally, I save both models using the pickle library and demonstrate how to use them in a Flask application for real-time predictions.
5: CNN- Developing a CNN model capable of accurately classifying sign language letters based on image inputs. By training neural networks on a dataset of sign languagegestures, the models will aim to recognize and classify hand gestures into their corresponding letters.
Dataset - The dataset consists of grayscale images of hand gestures representing sign language letters. Eachimage is of size 28X28 pixels, and the dataset includes labeled examples for training and testing.You may download the dataset from the google class page (hand sign.zip) and use it for the task.For the project, consider the Features: Input Image: Grayscale images of hand gestures representing sign language letters. Target Variable: The target variable is the class label corresponding to the sign language letter depicted in the input image.
LSTM - Implementing an LSTM-based deep learning model to forecast daily energy consumption for this household. Such forecasts could help in optimizing energy usage,reducing electric bills, and most importantly saving energy.
Dataset - Individual household’s electric power consumption data with a one-minute sampling rate over a periodof almost 4 years (source: UCI Machine Learning Repository). The dataset contains features like dateand time, power consumption (in kilowatts), voltage, current, power usage in different parts of the house etc.