hariprasath-v / DOCEREE_machine-learning-hackathon_round_1

Create a model that can accurately predict whether a user belongs to the HCP(Healthcare Professional) category or not. Based on server logs.

DOCEREE_machine-learning-hackathon_round_1

Leaderboard

• Rank:4
• Score:99.8824

Competition hosted on Techgig

Problem

Create a model that can accurately predict whether a user belongs to the HCP(Healthcare Professional) category or not. Based on server logs.

Evaluation

Evaluation metric for this competition is Accuracy.

Solution:

Exploratory Data Analysis

The basic exploratory data analysis of the data,

• Target distribution
• Categorical column level count

The above analysis had done by using,

• pandas
• numpy
• seaborn
• matplotlib

Model

Created catboost classifier model and tuned hyperparameters by using optuna framework. The model was evaluated by Accuracy.

Packages Used,

  * Sklearn
  * Pandas
  * Numpy
  * Matplotlib
  * catboost
  * optuna
  * shap

For more detailed information about the model.

Catboost Model Feature Importances

SHAP Catboost Model Feature Importances

Catboost Model train and validation accuracy

Catboost Model validation data confusion matrix

File information

doceree-machine-learning-hackathon-1-eda.ipynb

doceree-machine-learning-hackathon-1-model.ipynb