Name: Soham Kanti Bera
Type: User
Company: University of Stuttgart
Bio: Data Detective šµļø , Graduate Student @Uni_Stuttgart š§āš
Location: Stuttgart, Germany
Soham Kanti Bera's Projects
Learning the fundamentals of Bayesian Neural Networks (BNN) by doing.
A Cat and Dog image classifier using the pre-trained MobileNet V2 CNN Model of Google and Fine Tuning techniques
Using three supervised learning models for predicting whether a person is a potential donor for a charity organization and performing feature importance analysis to determine the most important features that impacts someone's probability of being a donor.
Various conceptual mini projects using C++ and Visual Studio
Finding clusters of similar demographic properties on general population data and then finding out which of the clusters are most likely to become a customer of the company.
Automatic classification of different severity grade [non-existent(0) to most severe(4)] of the Diabetic Retinopathy disease using deep learning
Classification of the chosen fashion product into one of the 10 fashion item categories using a neural network
Web app using Streamlit with a deep learning based back-end for food item name prediction from images and deploying the model in Heroku cloud.
creating multiple deep learning based image classifiers to identify dog breeds and comparing their performances to select the best model.
creating pipelines along with a Neural Network model for object detection in the CIFAR-10 dataset and comparing the model performance with a baseline
ETL and machine learning pipeline generation to categorize a data set containing real messages during disaster events for automatically sending to an appropriate disaster relief agency
convolutional neural network based detection and classification of objects using data from Waymo dataset
README file for my profile
Finding out various components like trends and seasonality in the time series describing tunnel traffic.
Building a logistic regression model for predicting passenger survival using the Titanic dataset