Competitive Intelligence: Dynamic Pricing

To solve real time problem of dynamic pricing based on Mercari-Price-Suggestion-Challenge on Kaggle using Deep Learning

Kaggle Competition

Pre-requisite:

1. Import the notebook on Google Colab
2. Download and unzip the training set (train.tsv.zip) from the above Kaggle competition link
3. Download the Word2Vec model Wiki.en.bin file
4. Create the similar folder structure (as given in instruction in 1st notebook) on Google Drive
5. Run below notebooks in the order in which they appear

We have 4 Jupyter Notebooks:

1. MPS_Data_Preprocessing_&_Wrangling.ipynb This notebook deals with data preprocessing and wrangling.
2. MPS_Entity_Embedding_Data_Modeling.ipynb This notebook deals with creating a DNN with data preparation for DNN like computing word2vec, training/evaluating the DNN model, training/evaluating the XgBoost model
3. MPS_Exploratory_Data_Analysis.ipynb This notebook performs exploratory data analysis. It needs some of the files created by the above two notebooks to do that.
4. MPS_RNN_Data_Modeling.ipynb This notebook performs data modelling using RNN

Results:

*EE = Entity Embeddings *DNN = Deep Neural Network