This machine learning project focused on predicting food delivery times. In this project, various essential steps were meticulously undertaken to predict food delivery times with precision. The journey began with thorough data cleaning to ensure the integrity of the dataset. Subsequently, feature engineering was employed to extract valuable insights, enriching the dataset. The application of label encoding and standardization further refined the data, enhancing the model's predictive capabilities.
For the regression task, the XGBoost algorithm took center stage, delivering an impressive R2 score of 0.82. This algorithm was not chosen arbitrarily; rather, it was the culmination of a comprehensive model comparison involving various regression algorithms. The selection process was facilitated by cross-validation, ensuring the model's robust performance across diverse datasets.
Throughout the development process, tools such as label encoding, standardization, and feature engineering were wielded as key instruments in crafting a predictive model that could meet the intricacies of the food delivery dataset. This repository encapsulates the synergy of meticulous data preparation, algorithmic selection, and model evaluation, providing a holistic view of the steps taken to achieve accurate predictions in the realm of food delivery timings.
This project used the "food-delivery-dataset/train.csv" sourced from Kaggle.
Link: https://www.kaggle.com/datasets/gauravmalik26/food-delivery-dataset?select=test.csv