Implemented a content-based Recommendation System which will understand the user Profile and based on that
recommend products to the user, in this case, product will be Movies!
Complete Data Pipeline has been used
in the implementation of project!
Used RFM and other approach involving calculation of various features which are average order value,
purchase frequency, profit margin, & churn rate to calculate the customer life time value.
Applied 9 different models (5 Deep Learning & 4 Machine Learning) to prepare a comparitive study!
Traffic Sign Classification
Used Le-Net Deep Learning Architecture to predict the traffic signs. Seperate training, calidatio,
& Testing data has been used, which leads to 99.93% accuracy of the Model!
Used K-Means clustering algorithm to classify the colours in the Image using predefined Number
of clusters, which are calculated using the elbow curve. Then a new image is constructed to show
the clustered colours in the image!
Classfying CIFAR-10 dataset!
Used Convolutional Neural Network to construct a model which will classify the CIFAR-10 dataset having
10 categories!
Applied Collaborative Filtering on the movie dataset, to find the most similar movie to a
given movie which can be recommended to a customer!
Fashion MNIST classification
Convolutional Neural Networks has been used in order to predict the image category of the image
present in the dataset!
Artificial Neural Networks has been used used to predict the price of a car which a person can
afford based on its various attributes!
Predicting Crime Rate in future
Facebook's Prophet is used to predict the crime rate in future for Chicago. It is a time series based
project which can be used for predicting future for any scenario!
Clustering Weather Stations
Various factors of weather stations are used to cluster the weather stations. This insight can be used for
implementation of strategies for particular weather for each weather station. Also, they can serve the purpose of
various other use cases!
