I’m Vishnu KP, and I have completed my B.E in Information Science & Engineering. I’ve been working as a Program Development Manager for Mansion Ecommerce for the past seven years. Just to try out the concepts learned in Great Learning, I decided to develop a “Ship Rate Prediction” Model that predicts ship rates for an item for a given country and zip code. Currently, we already have a Module that does the same, but it does with different TRY’s; when all try fails to get ship rate, then our current module initiates a shipping rates API to get the ship rates for a product. After developing the “Ship Rate Prediction” Model, we have decreased the shipping rate API call by 98%.
I have developed this model using LinearRegression, SVR, RandomForestRegressor, XGBRegressor, and TensorFlow Neural Network regression model. Out of the above-mentioned NN Regression Model, it gives good accuracy. So I used the NN Model for “Ship Rate Prediction.”
I relate the “Ship Rate Prediction” Model to the “Boston House Pricing” Model, which is one of the case studies explained in our Great Learning course videos. I have followed this case study and made changes according to my needs for building the “Ship Rate Prediction” Model. The current model is giving 95% accuracy against the test data. And in reality, the predicted Ship Rate is ± $0.50 to ± $0.95, which is acceptable.
Even though we started this to try out the concepts of AIML into our application just for curiosity after implementing this solution to one of our modules, we are amazed to see the efficiency and accuracy of this model. We have decreased the usage of Shipping Rates API during the Awards feed module. Now, Shipping Rates API is used only for the Real-Time Checkout processes.
Having said that, what started out of curiosity has opened gates to implementing AIML in our application, we have listed a few modules in our application where AIML can help in increasing the performance and better results.