The goal is to use the players’ attributes to predict his annual wage in dollars.
The evaluation is based on R-square of your predicted Wage vs the real value of Wage on the submission you made compared to the TA’s models.
We need to use the training data “FifaTrainNew” to build a valid model (include the process of data cleaning, graphing analysis, remove outliers/leverages or transformation) to predict the Wage based on the testing data “FifaNoY”.
The current R-square is around 0.9 in the attached Rmd file, and the output submission files should contain two columns: Ob and WageNew. with a csv format (like the attached “newWage.csv”) .
Please try to make adjustments on the attached Rmd to improve R-squared (we need to arrive at least 0.958) and we will test the R-squared on Kaggle for your output file for testing.
0 comments