As public awareness of pet welfare continues to
increase, pet owners face growing challenges in identifying reliable and
high-quality pet service providers. Existing selection processes often rely on
fragmented online information and subjective judgments, resulting in suboptimal
decision-making. To address this issue, this study proposes a deep
learning–based evaluation and recommendation framework that integrates recurrent
neural networks (RNNs) with web crawling techniques to support systematic and
evidence-based pet service selection. The proposed framework utilizes the
official government dataset of licensed pet businesses as a foundational data
source and enriches it with large-scale consumer reviews and supplementary
attributes collected from major social media platforms and online review
websites. Key decision factors influencing pet owners’ choices are identified
and incorporated into the model training process. By leveraging the sequential
learning capability of RNNs, the framework captures individual user preferences
and generates ranked evaluations and personalized recommendations for pet
services.
The effectiveness of
the proposed approach is demonstrated through an Android-based application that
delivers tailored recommendations to end users. The results indicate that the
framework enhances recommendation accuracy and improves user decision satisfaction.
Beyond assisting individual pet owners, this study contributes an objective and
scalable decision support framework that promotes service quality improvement
and transparency within the pet service industry.
