Using various regressors to predict used car prices in india.
- Engineered important features that impact the price of used car (brand, mileage, speed).
- Utilized numerous regression models to obtain the highest performance.
- Accuracy was measured if the predicted value was within a certain range of the actual value of a given car.
- Decision tree regression yielded a accuracy of 76% (76% of predictions were within 3000$ of their actual value).
Below is an image showing the residuals of our decision tree regressor
A more detailed presentation can be seen in the files as well as the EDA notebook