TITLE:
Predicting Adult Dog Temperament Based on Puppy Behaviors: A Machine Learning Approach for Enhancing Canine Welfare
AUTHORS:
Ashen Mihiranga Benthota Pathirana, Himendra Balalle
KEYWORDS:
Dog Temperament Prediction, Canine Welfare, Puppy Behaviour, Machine Learning, Random Forest, Responsible Dog Ownership
JOURNAL NAME:
Open Journal of Applied Sciences,
Vol.14 No.11,
November
8,
2024
ABSTRACT: Every year, a higher number of dogs are abandoned or euthanised due to temperament issues and a lack of understanding by owners regarding dog behaviour and training. This research focuses on the potential to make predictions of adult dog temperament based on early puppy behaviours by using a machine learning model. Specifically, the research used guard dog breeds, such as American Bully, American Pit Bull Terrier, and German Shepherd. The study collected dog data and general data from dog owners and used the Random Forest approach to build a predictive model. Users are allowed to input puppy data and receive adult dog temperament predictions in model, which is integrated into a web application. The aims of this web application are to enhance responsible dog ownership and reduce abandonment by offering insights and training recommendations based on predicted outcomes. The model achieved a prediction accuracy of 86% on testing, and it is continually improving, though further refinement is recommended to improve its reliability and applicability across a broader range of breeds. The study contributes to canine welfare by providing a practical solution for predicting temperament outcomes, ultimately helping to reduce shelter populations and euthanasia rates.