SAS Analytics Day Using SAS® Enterprise MinerTM to Predict the Number of Rings on Abalone Shells Gangarajula Ganesh, SAS & OSU Data Mining Student, Oklahoma State University Seetharama Yogananda Domlur, Manager II – Customer Analytics, Walmart 4/28/2015 1 Introduction – What is an Abalone? Why Analyze it? A type of sea snail with shell covering. Layered rings form the shell, which grow with age Shell structure under research to build strong body armor. Traditional approach of counting rings is tedious and time consuming. 4/28/2015 2 Data Description Data Source: UC Irvine Database. 4177 Observations, 10 independent variables 4/28/2015 Variable Name Description Length Length of the shell Diameter Diameter of the shell perpendicular to length Height Height of the shell Shell Weight Shell weight after being dried Shucked Weight Weight of the meat Viscera Weight Gut weight after bleeding Whole Weight Whole Abalone weight Male Male indicator Female Female indicator Infant Infant indicator Rings Number of rings on the shell 3 Data Preparation Data Partitioning • Training – 70% • Validation – 30% Summary Statistics 4/28/2015 4 Modeling 4/28/2015 5 Results Best selected model: Neural Network Best explainable model: Stepwise Linear Regression Selection Criteria: Average Squared Error 4/28/2015 6 Model Results: Stepwise Linear Regression Model Parameter Estimates: 4/28/2015 7 Conclusions 1. The number of rings increase by 10.78 units for a 2. 3. 4. 5. unit increase in Diameter. The number of rings increase by 9.41 units for a unit increase in Height. The number of rings increase by 19.27 units for a unit increase in Shell Weight. The number of rings decrease by 0.63 units for a unit increase in Infant class, signifying an infant. The number of rings increase by 0.23 units for a unit increase in Gender‐Female class. 4/28/2015 8 Acknowledgement We thank Dr. Goutam Chakraborty, Ralph A. and Peggy A. Brenneman Professor of Marketing & Founder of SAS and OSU Data Mining Certificate Program‐Oklahoma State University for all his support and guidance throughout the project. 4/28/2015 9 References http://www.aps.org/publications/capitolhillqu arterly/201105/seasnails.cfm http://archive.ics.uci.edu/ml/datasets/Abalon e [Bache, K. & Lichman, M. (2013). UCI Machine Learning Repository [http://archive.ics.uci.edu/ml]. Irvine, CA] 4/28/2015 10 Disclaimer We have analysed this topic using standard data mining and statistical techniques and we do not claim to have any kind of expertise in understanding the biology of Abalone Shells. 4/28/2015 11 Thank You!! Ganesh Gangarajula [email protected] https://www.linkedin.com/in/gangarajulaganesh Yogananda Domlur Seetharama [email protected] https://www.linkedin.com/in/domlur 4/28/2015 12
© Copyright 2024