Regression Analysis
Quantitative Methods Project
Regression Analysis for the pricing of players in the Indian Premier League
Executive Summary
The selling price of players at IPL auction is affected by more than one factor. Most of these factors affect each other and still others impact the selling price only indirectly. The challenge of performing a multiple regression analysis on more than 25 independent variables where a clear relationship cannot be obtained is to form the regression model as carefully as possible.
Of the various factors available we have leveraged SPSS software for running our regression analysis. One of the reasons for preferring SPSS over others was the ease with which we can eliminate extraneous independent variables. The two methodologies used for choosing the best model in this project are: * Forward Model Building:
Independent variables in order of their significance are incrementally added to the model till we achieve the optimum model. * Backward Elimination:
The complete set of independent variables is regressed and the least significant predictors are eliminated in order to arrive at the optimum model.
Our analysis has shown that the following variables are the most significant predictors of the selling price:
COUNTRY : whether the player is of Indian origin or not
AGE_1 : whether the player is below 25 years or not
T_RUNS : total number of test runs scored by the player
ODI_RUNS : total number of runs scored in ODI matches
ODI_WICKET : total number of wickets taken by the player
RUNS_S : total number of runs scored by the player
BASE_PRICE : the base price of the player set in IPL
Using the calculated coefficients the regression model equation can be stated as below:
SOLD PRICE = -13366.247 + 219850.349(COUNTRY) + 204492.531(AGE_1) -59.957 (T_RUNS) + 53.878
Regression Analysis for the pricing of players in the Indian Premier League
Executive Summary
The selling price of players at IPL auction is affected by more than one factor. Most of these factors affect each other and still others impact the selling price only indirectly. The challenge of performing a multiple regression analysis on more than 25 independent variables where a clear relationship cannot be obtained is to form the regression model as carefully as possible.
Of the various factors available we have leveraged SPSS software for running our regression analysis. One of the reasons for preferring SPSS over others was the ease with which we can eliminate extraneous independent variables. The two methodologies used for choosing the best model in this project are: * Forward Model Building:
Independent variables in order of their significance are incrementally added to the model till we achieve the optimum model. * Backward Elimination:
The complete set of independent variables is regressed and the least significant predictors are eliminated in order to arrive at the optimum model.
Our analysis has shown that the following variables are the most significant predictors of the selling price:
COUNTRY : whether the player is of Indian origin or not
AGE_1 : whether the player is below 25 years or not
T_RUNS : total number of test runs scored by the player
ODI_RUNS : total number of runs scored in ODI matches
ODI_WICKET : total number of wickets taken by the player
RUNS_S : total number of runs scored by the player
BASE_PRICE : the base price of the player set in IPL
Using the calculated coefficients the regression model equation can be stated as below:
SOLD PRICE = -13366.247 + 219850.349(COUNTRY) + 204492.531(AGE_1) -59.957 (T_RUNS) + 53.878