nikhitamethwani Goto Github PK

adm_project_v1

TEST Project

amazon-alexa-review-using-sentiment-analysis

Sentiment Analysis on the text data to predict the ratings by the users on the Amazon products

bayesian-inference---mlb-batting-prediction

Predicting Batting order of Red Sox against Yankees using Bayesian Inference

big-data-technology

Analysis on Amazon Health care Products using Big Data Technologies using HDFS , Map/Reduce , Hive , PIG

black-friday-sales-prediction

Data Analysis on Predicting the sales on Black Friday

cnn_hyperparametertuning

This repository discusses about training many Convolution Neural Networks and studying the effect of training the network by tuning hyper parameters such as network architecture,filter size, optimizers,activation functions and determining the best model with high testing accuracy and less training loss

csye_7245

CSYE 7245 - Big-Data Systems and Intelligence Analytics

data-analysis-mortality-age-death-specific-to-country-tableau

Data Analysis and Visualization is done through Tableau for the Mortality age death specific to country.Dashboard is created which provides filtering effect and allows user to check the Average Deaths as per Country and Sex.

data-analysis-on-insta-cart-

database-management-system

Designed E commerce Database & implemented Stored Procedures ,Triggers , views on the tables

deep-fakes-cars-generation

In this project , We introduce understanding of GANS for generating fake images. Its network architecture and problem associated with it are covered.We move to Progressive GANS which are found to be better in producing high quality images and introduce the working of Porgressive GANS architecture introduced in Research Paper https://arxiv.org/abs/1710.10196 The main aim of this project was to research about the working of Progressive GANS on Celeb dataset,understand the functionality of layers and building blocks of the network and reimplement it on car dataset. The heavy training time of GANS led to the introduction of Google Cloud Platform which is used to train the network to generate the fake cars. The input dataset was preprocessed to 128 * 128 size for training the network The reimplementation of standard code is done by changing the required code as per the car dataset specifcation which is explained further in details in the notebook. The network is trained for 12 hours to generte fake low resolution car images. The resolution of fake images can be improved further by increasing the training time of the network

deep-learning---mlp-hyperparametertuning

This repository covers the concepts of MLP and effect on the model by tuning various hyperparameters of the network.Multiple permutation combination of network tuning are done to determine the best model having highest test accuracy with no overfittin/underfiting problems.

deep-learning-cnn---cifar-hyper-parameter-tuning

The repository uses the CIFAR-10 Datasets and tests various models by hyper parameter tuning its layer architecture,Batch Normalization,optimizers and understands the factors related to overfitting and underfitting with these models

deeplearning_andrewng

Exercises for Andrew Ng's Deep Learning Coursera

generative-adversarial-networks

Introduction to generative adversarial networks, with code to accompany the O'Reilly tutorial on GANs

good-books-read-recommendation-pipeline

Developed Flask App to recommend books using data from AWS S3 Data storage and deploying the app on cloud

house-price-predicition

Regression Problem having the requirement of predicting the sales price of Houses. Data Engineering,Feature Extraction , Visualization was done on the data to fetch the features having alligned relationship with the predictor variables.GridSearchCv for hyperparameter tuning and Pipelining for setting up the process workflow was used to train the models and evaluate each models on the basis of accuracy

info7374-digital-marketing

Digital Marketing

installation-guide

Installation Guide Docs

rossman-sales-prediction---time-series

Forecasting Sales of next 6 weeks of Rossmann store USING SARIMAX model which allowed the seasonal effects into consideration.The accuracy was improved further by using XGBoost by training it for 6000 iterations.

sentiment-analysis