Abstract
Determining whether or not the indexed price of a vehicle has increased or decreased is a difficult undertaking because of the numerous factors that influence a used car's market value. This project intends to develop machine learning models that can exactly anticipate a used car's price depending on its features, helping customers to make intelligent conclusions. The supervised system learning approach is used in this research to build a system for predicting car prices. The investigations employ a few linear regressions as part of the system learning and prediction approach, which provided 98% prediction accuracy. Multiple unbiased variables are employed in multiple linear regressions, the true and predicted values of only one dependent variable, yet, are compared in order to assess the accuracy of the effects. The knowledge came from classified ads for automobiles used. Regression approaches of all kinds were employed by the researchers, the researchers used a range of regression methods, such as desire tree regression, linear regression, manual vector regression, and random regression, woodland regression, to get the highest level of accuracy. This research paper's objective is to examine whether auto charges correspond to the information provided (previous customer information like engine capability, distance travelled of manufacture, and so forth). After that, it was decided which method best suited the available fact set by comparing the relative performances of the several algorithms. Right now, a Java utility makes use of the earlier prediction model. Additionally, test data were used to assess the model's accuracy performance, and a result of 87.38% was obtained.
Keyword
multiple Vehicle rate, regression version, linear regression machine learning, Correlation
PDF Download (click here)
|