Abalone-Age-Prediction

University of California, Davis STA206 project

Language being used: R studio

Abalones are well known for their rarity and delicacy. Abalones are expensive and their prices are related to their ages. This project is interested in developing predictive models for estimating abalone age with higher accuracy, finding the model with best performance, and obtaining the most significant indicators for prediction. A comprehensive data preprocessing is done to tidy the data set, including handling missing values and outliers and train-test split. Exploratory data analysis is conducted to better understand the data, such as the distributions and correlations. After that, four regression models, including multiple linear regression, forward stepwise regression, LASSO regression, and random forest are built, validated, and evaluated. As a result, the stepwise regression, LASSO regression, and random forest models have their own advantages and limitations. Also, further discussions are made for the models, data set, and the entire analysis process at the end of this project.