Implementing Ensemble techniques to predict license status for the given business.
The dataset used is a licensed dataset. It contains information about 86K different businesses over various features. The target variable is the status of the license, which has five different categories.
- Language:
Python
- Libraries:
pandas
,scikit_learn
,category_encoders
,numpy
,os
,seaborn
,matplotlib
,hyperopt
,xgboost
- Data Description
- Exploratory Data Analysis
- Data Cleaning
- a. Missing Value imputation
- b. Outlier Detection
- Data Imbalance
- Data Encoding
- Model Building
- a. Random Forest
- b. AdaBoost
- c. XGBoost
- Feature importance
- Hyperparameter tuning
- a. Random search optimization
- b. Grid search optimization
- c. Bayesian optimization
-
The
input
folder contains the data that we have for analysis. In our case, it containsLicence_Data.csv
. -
The
src
folder is the heart of the project. This folder contains all the modularized code for all the above steps in a modularized manner.The
model_selection.py
andpreprocessing.py
files contain all the functions in a modularized manner, which are appropriately named. These Python functions are then called inside therun.py
file.The
requirements.txt
file has all the required libraries with respective versions. Kindly install the file by using the commandpip install -r requirements.txt
. -
The
output
folder contains an Excel file for classification metrics scores of each model. -
The
lib
folder is a reference folder. It contains the original Jupyter notebook.