hariprasath-v / Hackerearth-amazon-business-research-analyst-hiring-challenge

Build a machine learning model that can calculate the time the delivery person takes to deliver the order.

Hackerearth-amazon-business-research-analyst-hiring-challenge

Competition hosted on hackerearth.com

About

Build a machine learning model that can calculate the time the delivery person takes to deliver the order.

Final Score is 82.15

Evaluation Metric 100*R2 score.

File information

• hackerearth-amazon-data-preparation.ipynb

  The dataset is in a little bit tricky format. Each and every row of the train and test dataset is in a separate text file. To train a model have to create a data frame from the text file information.

• hackerearth-amazon-hiring-eda.ipynb

  Basic Exploratory Data Analysis

  Packages Used,

   * seaborn
   * Pandas
   * Numpy
   * Matplotlib
  

• hackerearth-amazon-hiring-model.ipynb

  Data Pre-processing and model.

  Packages Used,

    * Sklearn
    * Pandas
    * Numpy
    * Matplotlib
    * pycaret    
  

  Compared multiple regression models using pycaret’s compare_models function. Then took the top 3 models based on the r2 score then blend the model by using pycaret blend_models function.

  For more detailed information about the model.

Lightgbm Regressor Residual Plot

Lightgbm Prediction Error Plot

Top 3 Models

Lightgbm Feature Importance Plot

SHAP - Lightgbm Feature Importance Plot

Time_taken_(min) distribution of train and test data