Track: High School STEM Project Presentation
Abstract
This STEM poster is to build a statistical model to predict the team winning percent based on their offensive, defensive, and differential statistics during 2015-2016 NBA regular seasons. The team statistics have been standardized to Z score in each category to remove any mean and standard deviation effect. We used a multiple linear regression model to predict the team winning record. After trimmed the insignificant input variables based on the significance P-Values, the predictive model can predict team winning percent with R-Square > 0.95. The regression model has indicated that the importance of 3-point Percentage, Turn Over, and Point Per Game are critical to the offensive efficiency. In defense, how to defend the rebound and opponent’s field goal percentage are most critical. Warriors’ team record has been identified as an outlier since their winning percent and team statistics are separated from the other teams. We have also considered the 2nd-order Interaction Term. Defense Field Gold% * Defense Point Per Game is the most significant interaction term. This model can predict whether Warriors can break their 73-win record. This model may also provide NBA coaches how to build a better team to win more games.