The model explains about 43.6% of the variance in car prices, with a Root Mean Squared Error (RMSE) of approximately 8652 and a Mean Absolute Error (MAE) of around $6134. Among all variables, mileage, brand, and state were the most important predictors. While state population had some influence, it was less significant compared to other features. This suggests that car prices are affected more by vehicle-specific details than by population size alone.
Embedded Files
Mileage is the most important feature — the model relies on it heavily.
Brand, state, and year are also important.
Population has the least impact, which suggests population size in a state doesn’t strongly predict car prices on its own.
The decision tree model helps visualize how car price is influenced by key features. The first and most important split is based on mileage. Cars with higher mileage tend to have lower prices. For example, vehicles with mileage above 29,504 miles generally had an average price around $15,387, while those with lower mileage had a higher average price of $25,780.
The tree further splits based on mileage again and then brand. For instance, if a car’s mileage exceeds 77,334 miles, its price drops to around $9,734. On the other hand, certain brands like Honda, Kia, and Nissan are associated with different price behaviors, which shows how brand identity also plays a role.
Overall, this tree highlights how mileage and brand are strong predictors of used car prices.
This plot shows how well the model predicted car prices compared to the actual prices. Each circle represents one car listing. The red line is where perfect predictions would land meaning the predicted price exactly matches the actual price.
Most of the dots are pretty close to that line, especially for cars priced under $30,000, which means the model is doing a decent job overall. However, for some higher-priced cars, the predictions are a bit off usually underestimating the value. This could be because there’s less data for expensive cars, or the model might be missing some features that affect luxury pricing.Overall, the model captures the pricing trend well, but there's still room to improve especially for the more expensive vehicles.
CODE
Page updated
Google Sites
Report abuse