Soumayan Pal's Projects
All the Tasks and Assignments submissions done in the AI ML Training Program of KIIT CAAS are present here.
In this project we have performed a detailed analysis and visualization of the training dataset with different Exploratory Data Analysis techniques. Then a comaparative analysis of different Classifier models have been done to predict the passenegr satisfaction class. After performing the model comparative analysis we can conclude that the Random Forest Classifier and the Support Vector classifier have performed better. HYpertuning on both of these classifiers have been done to reduce overfitting and achieve an improved accuracy of 97.15% and 96.57% respectively.
To predict and classify the levels of Airlines Catastrophes using classification algorithms.
The aim of this project is to recognize what the person is trying to convey using different hand gestures. The dataset contains 29 classes which comprises of A to Z alphabets, nothing, space and delete hand gestures. Here first we will first try training a Multi Layer Perceptron model and check its accuracy to recognize the gesture labels. Then we have used the CNN approach.
The aim of the project is to make analysis and forecast of the demand in Bike Rental Services. The dataset contains 2 years data different features that are responsible for fluctuations in the customer ride request demand for training the model. This project predicts the upcoming nature of the customer request demand.
A simple Birthday Card developed in Android Studio
A guide to creating a chatbot with Rasa stack and Python and deploying it on Slack
The aim of the project is to predict fraudulent credit card transactions using machine learning models.This analysis and prediction is important for a bank as each fraud trancastion is a loss of the bank as well as customer faith. The dataset contains transactions made by credit cardholders.
An app with a Button to roll the dice which gives a randomly generated number between 0 to 6, each time its rolled.
The aim of this project is to analyse the reviews given by visitors from different countries of the world using NLP to understand the sentiment of the reviews and classify using Sentiment Analysis metrics like Sentiment Polarity and VADER Polarity. This processed data is then feeded to different classifier models to get trained and predict the sentiment of the test reviews.
The aim of this project is to recognize the computerised generated images of the English apphabets. The dataset contains 26 classes which comprises of a to z alphabets, each class containing 100 images.In this project we have used a couple of Deep Learning frameworks to classify the computerized images of English alphabets.
A real time Face detection model that detects real time multiple faces in the device's camera using OpenCV.
This project is about building a model which detects faces with mask and those without mask in real time using Caffe framework of CNN. The Caffe model takes the real time video as input to detect whether the face detected, is wearing a mask or not and put an appropriate message.
This project is about building a model which detects faces with mask and those without mask in real time using OpenCV. The Haar Cascades Classifier is being here to detect whether the face detected, is wearing a mask or not and put an appropriate message.
The aim of the project is to detect whether the news is Real or Fake using different text extraction NLP techniques to understand the data, use different classifiers approaches and RNN framework on this training data to train various models and use them to make detections.
A Hypertuned Random Forest Regressor model that predicts flight fares.
In this project we have performed a detailed analysis and visualization of the training dataset with different Exploratory Data Analysis techniques. Then a comaprative analysis of different Regressons have been done to predict the flight fare predictions.After performing the model comparative analysis we can conclude that the hypertuned Random Forest regressor model performs best with an accuracy of 84%
The aim of this project is to predict the price of gold
Predict gold price using Machine Learning
Full Machine learning Guide basic to advance
This repository contains mini projects in machine learning with notebook files
A project to build a Facial recognition model using Support Vector Machines classifier to predict the faces and names of the lfw_people dataset from scikit learn.
Images compression is done by converting the pixels in numpy arrays and then using Kmeans Clustering classification algorithm to compress all colors to selected 'k' colors. An interactive control is also setup fro direct access of images from directory and using slider to denote 'k' value from 1 to 256
Performing Linear Regression on random numbers using Gradient Descent algorithm.