Product Price Recommendation Using Product Descriptions
To build an algorithm that automatically suggests the right product prices based on the user provided textual product descriptions,
and other details like product category, brand name, and item condition.
The Highlights:
- Used Regresstion Tree Model with Gradient Boosting as the model
- Performed hyper parameter tuning using regressor techniques in LGBM library
- RMSE 0.411,Top 100 Kaggle Public leaderboard
This model can be used to classify a given image of a handwritten digit into one of 10 classes representing integer values from 0 to 9, inclusively.
I have tested 2 different machine learning models(SVM and CNN) to find the best model for the classification task. After the selection of the model(VGG CNN architecture), I have applied some hyper-parameter tuning to create the best possible model in terms of accuracy of the digit classification on the test dataset.
Created a collaborative filtering recommendation system by discovering the preferences of users from the data, creating a utility matrix consisting of each user-item rating values and predicting the blank cells in this utility matrix.
The Highlights:
- Used Vector Space Modeling(Cosine Similarity) to find similar users.
- Performed matrix factorization using Low rank factorization.
- Used Adam’s optimization algorithm for minimizing the cost function.
Created a Seasonal ARIMA Time series model to forecast the price of Bitcoin.
The Highlights:
- Performed EDA to get insights and differencing orders of the model.
- Created an application on RShiny.
An analytical report(Using Tableau) to improve the cab ridership in New york for the Taxi and Limousine Commission(TLC)
Our Recommendations:
- Introduction of Shared Rides
- Tapping Dormant Boroughs
- Price Prediction Application(using xgBoost)
