Posts

Showing posts from January, 2019

Adventures with R - Cricket Analysis (Predicting performance of Players based on previous ODI performances)

This is a continuing series on R and work on Cricket Analysis Previous Series Basic Setup Getting all players in a single file Clustering players based on ODI performance This blogpost is focused on trying to predict individual player's performance on Runs the player will score based on performance in previous matches the player has played ** This code took longer to set up. The initial work was testing out different regression techniques, I kept coming back to the basic logistic regression as the model performance was not that improved between the runs. I believe the data is thin at an individual player level and therefore most regression techniques are having an issue with respecting to predicting runs scored. The model itself is not that great, it is basically predicting runs the player will score based on the balls faced, the Venue Type (Home or Away) and the Opposition Type (Strong or Weak) Files are available at Code Full code is reproduced for reference li