With the ongoing demand for data analysis and prediction, machine learning has become a pivotal area of research and experiment. This project outlines a data-driven approach to predict the outcomes of shelter dogs and cats, aiming to enhance resource allocation, improve animal welfare, and reduce euthanasia rates. While most of the state-of-the-art approaches prioritized leveraging a single machine learn- ing algorithm, our research emphasizes integrating various features including demographic information, behavioral traits, and health status which enables the proposed framework to improve upon existing predictive accuracy. The project will initially focus on a localized open dataset from the City of Austin. Predictive analysis is performed using Naive Bayes, K-nearest Neighbours, Classification Tree, and Random Forest. Evaluation will be conducted using historical shelter data, with accuracy, precision, and recall metrics. By providing shelters and rescues with actionable insights, the project aims to facilitate quicker outcomes for animals, strengthen partnerships with fosters, and ultimately improve the well-being of shelter animals across the country.
-
Clone the git repository to your machine (makes a copy).
-
To set up the project environment, make sure you have Anaconda/Conda installed, then open the command prompt on your machine.
-
Use the
cdcommand in the command prompt to navigate to where you cloned the repository (where the files for the project are located on your computer). -
Now, when working on the project, you'll have an environment that is exactly the same as ours, meaning all the packages are exactly the same.
The datasets we used can be found at https://data.austintexas.gov/browse?q=animal.