This dataset contains a collection of more than 19,000 how-to guides publicly available on wikiHow. The data is extracted using the open-source API wikihowunofficialapi. The extracted text has been scored by the wikiHow users on a scale of 0 to 100, representing their helpfulness. The model will be used to predict which article will score high or low based on our features. I have created a Pipeline for each of these models, such that the data transformations and model training/prediction steps are assembled together.
The prediction is a regression problem, so I have trained three models: LinearRegression, RandomForestRegressor and XGBRegressor. I have done EDA.