Pinned Repositories
Car-Sales-Prediction
Developed a Sequential model, implemented linear regression using Artificial Neural Network to predict the total purchase amount in dollars the customers are willing to pay given the attributes like age, salary, net worth, debt, etc.
Digit_Recognizer
Built an artificial neural network for digit recognition from scratch applying the feedforward and backpropagation algorithms using numpy. • Used sigmoid function for hidden layer followed by softmax function with cross entropy function for output layer to get an accuracy of 87.4%.
FIFA2019_player_classification_prediction
Implemented machine learning algorithms to predict and classify FIFA 2019 players and their overall rating. Used Linear & Random Forest Regression models to predict the overall rating of the players based on various attributes like sprint speed, shot accuracy, stamina, etc. With Naive Bayes Classifier delivered the classification of players w.r.t to their positions on the field.
Instagram-Clone
Developed a clone of Instagram which allows users to make their own accounts to share pictures and thoughts over the internet. Constructed it using Amazon Web Services’ EC2 instance along with a Parse Server, a database, to store information about users and the images they share. Allows users to sign up, log in, and stay logged in or log out. An app that allows users to make their own accounts to share pictures and thoughts over the internet.
Music-Recommendation-System
Built a music recommendation system using observable data of users on KKBOX music streaming service with various factors of songs, members, and other processed information based on the user activity using XGBoost & Random Forest Classifiers.
RecruTech
The project is used for the recruiters which can give the input of particular job position.The project will search all the potential candidates and based on recruiter's choice he/she can filter out using skills. Used BeautifulSoup 4 and Selenium libraries for parsing HTML content from the site. Implemented using XGB & Random Forest Classifier. Used word2vec for vectorizing job profile data.
AvocadoMarket
Predicting future prices of Avocado in the US using Facebook Prophet for time series considering multiple attributes like year, type, size, region, volume of the commodity, etc. Data represented is the weekly 2018 retail scan data for National retail volume (units) and price.
Heart-Disease-Prediction
Using Supervised Machine Learning algorithms like Linear Regression, SVM, KNN, Naive Bayes & Random Forest presence of a potential cardiac disease in a person w.r.t. factors like age, heart rate, ST depression level, serum cholestrol, fbs. etc. is predicted using integer value from 0 (no presence) to 4. Delivering test accuracy of 85% using Linear Regression, Naive Bayes Classifier & Random Forest algorithms.
Image_Classifier
Built a CNN model to perform image classification using Keras on the cifar10 dataset. Performed 2D convolutions, MaxPooling2D, AveragePooling2D. Optimized the network weights using Adam Optimizer. Also, performed network regularization techniques such as Dropout. Evaluated the model, using confusion matrices and classification reports. For enhancing the network generalization capability, did image augmentation.
Twitter-Clone
A clone of the app that allows to people to share their thoughts, ideas and views on a social networking platform.
harshvasa's Repositories
harshvasa/2D-MatrixTranspose
Python Code for transposing a 2D-matrix
harshvasa/Automobile-Price-Prediction
Used Azure Machine Learning Studio for building quick and effective Machine Learning models. In my model I used Linear Regression to predict the price of vehicles from the Automobile price dataset.
harshvasa/Malaira-Detection
Created a Sequential Model, demonstrated whether the image is infected or not from the given dataset, Used adam optimizer algorithm, convo2D & 2D pooling techniques. Using ImageDataGenerator tested random samples to see the output, which after multiple epochs resulted in an accuracy of 94.07% and validation accuracy of 95.29%.
harshvasa/Image_Classifier
Built a CNN model to perform image classification using Keras on the cifar10 dataset. Performed 2D convolutions, MaxPooling2D, AveragePooling2D. Optimized the network weights using Adam Optimizer. Also, performed network regularization techniques such as Dropout. Evaluated the model, using confusion matrices and classification reports. For enhancing the network generalization capability, did image augmentation.
harshvasa/AvocadoMarket
Predicting future prices of Avocado in the US using Facebook Prophet for time series considering multiple attributes like year, type, size, region, volume of the commodity, etc. Data represented is the weekly 2018 retail scan data for National retail volume (units) and price.
harshvasa/Chicago-Crime-Rate-Prediction
Applied Facebook Prophet for Time Series Prediction of the crime rate in Chicago using the dataset acquired by the CLEAR System of the police department from 2001 to 2017. The dataset contains various factors that include every detail about the crimes reported. Performed data visualization on crime types and resampled the data for time series. Predicted the crime rate and plotted a graph that forecasts the trend in crime rates, monthly, quarterly, annually, etc.
harshvasa/Breast-Cancer-Detection
Implemented Supervised Machine Learning Classifiers like XGB, SVC, Naive Bayes. Random Forest on a dataset that consists of various diagnostic measures divided into three classes of se, mean and worsXGB and Random Forest classifiers are the best-fit with an accuracy of over 95% each. Implemented Supervised Machine Learning Classifiers like XGB, SVC, Naive Bayes. Random Forest on a dataset that consists of various diagnostic measures divided into three classes of se, mean and worst. XGB and Random Forest classifiers are the best-fit with an accuracy of over 95% each.
harshvasa/Bitcoin_Price_Prediction
The cryptocurrency price prediction is estimated by training and testing Random Forest Regressor algorithm on the model based on the data generated of the actual bitcoin price quotation from August 2018 to 2019. The prediction of prices is carried out for 1 month after the end date.
harshvasa/Diabetes_Prediction
Made use of Supervised Classifiers like SVC, KNN, Naive Bayes, Random Forest, XGB to predict the presence of diabetes in the patient on basis of the predictor variables like the number of pregnancies the patient has had, their BMI, insulin level, age, etc. XGB Classifier resulted in giving the most accuracy of 81.82%.
harshvasa/Flick_Soccer
A free-kick battle on the football field as the user tries to swipe and flick the balls into the net with the controls added. It was developed using Unity. Special Trail effect is added to the ball.
harshvasa/Fruit_Ninja
Developed the clone of the actual game fruit ninja using a touchpad to slice on-screen fruits with the Unity3D engine. Additional points are rewarded for slicing multiple fruits in one swipe.
harshvasa/Image_Data_Augmentation
Learned how to apply image data augmentation using keras. With ImageDataGenerator class from Keras’ image preprocessing package, dealt with a variety of options available in this class for data augmentation and data normalization.
harshvasa/Customer-Billing-System
A user-generated billing system that lets the user add the account details to the server and then search by name or the account number. The payment option is available that will compute the status of the holder by checking with balance and showing response as added (Delinquent/Overdue). Finally the bill will be printed with the current status of the user.
harshvasa/Music-Recommendation-System
Built a music recommendation system using observable data of users on KKBOX music streaming service with various factors of songs, members, and other processed information based on the user activity using XGBoost & Random Forest Classifiers.
harshvasa/Heart-Disease-Prediction
Using Supervised Machine Learning algorithms like Linear Regression, SVM, KNN, Naive Bayes & Random Forest presence of a potential cardiac disease in a person w.r.t. factors like age, heart rate, ST depression level, serum cholestrol, fbs. etc. is predicted using integer value from 0 (no presence) to 4. Delivering test accuracy of 85% using Linear Regression, Naive Bayes Classifier & Random Forest algorithms.
harshvasa/Car-Sales-Prediction
Developed a Sequential model, implemented linear regression using Artificial Neural Network to predict the total purchase amount in dollars the customers are willing to pay given the attributes like age, salary, net worth, debt, etc.
harshvasa/ZigZag_3D_game
The main objective of this game is to roll the ball through the path. This game is a clone version of original Zig Zag. The players have to follow the path by rolling the ball without falling down. All the game controls are under a single mouse click. It is a 3D game filled with sharp twists and turns developed and designed using Unity 3D and C#.
harshvasa/RecruTech
The project is used for the recruiters which can give the input of particular job position.The project will search all the potential candidates and based on recruiter's choice he/she can filter out using skills. Used BeautifulSoup 4 and Selenium libraries for parsing HTML content from the site. Implemented using XGB & Random Forest Classifier. Used word2vec for vectorizing job profile data.
harshvasa/Digit_Recognizer
Built an artificial neural network for digit recognition from scratch applying the feedforward and backpropagation algorithms using numpy. • Used sigmoid function for hidden layer followed by softmax function with cross entropy function for output layer to get an accuracy of 87.4%.
harshvasa/FunkyFashionista-Online_Fashion_Store
An online multi-faceted fashion store that allows customers to purchase clothes from existing range of brands.
harshvasa/InfiniteRunner-3D_Unity
Infinite 3D Runner collecting candies & avoiding obstacles.
harshvasa/FIFA2019_player_classification_prediction
Implemented machine learning algorithms to predict and classify FIFA 2019 players and their overall rating. Used Linear & Random Forest Regression models to predict the overall rating of the players based on various attributes like sprint speed, shot accuracy, stamina, etc. With Naive Bayes Classifier delivered the classification of players w.r.t to their positions on the field.
harshvasa/SuperMario-Run-Clone
A clone of the game, Super Mario in which the user is Mario himself and runs along the tracks collecting coins, avoiding bombs
harshvasa/Uber-Clone
A clone of the app which allows riders and drivers to connect using anonymous login and locations. It has been implemented using multiple customizations for riders and drivers using Google Maps Activity.
harshvasa/Whatsapp-Clone
A clone of the app, a cross-platform instant messaging, multifaceted application. Allows users to send and delete messages sent to their friends.
harshvasa/Twitter-Clone
A clone of the app that allows to people to share their thoughts, ideas and views on a social networking platform.
harshvasa/Flappy-Bird-Clone
A clone of the game, Flappy Bird in which the user plays by tapping on the screen trying to move through all pipes without collision while saving the bird from falling out of the screen.
harshvasa/Instagram-Clone
Developed a clone of Instagram which allows users to make their own accounts to share pictures and thoughts over the internet. Constructed it using Amazon Web Services’ EC2 instance along with a Parse Server, a database, to store information about users and the images they share. Allows users to sign up, log in, and stay logged in or log out. An app that allows users to make their own accounts to share pictures and thoughts over the internet.
harshvasa/Notes-App
An app that lets the user create and customize their important notes they make.
harshvasa/NewsReader-App
An app to show a list of news headlines and the corresponding news articles. (Online and Offline)