The rapid advancement of artificial intelligence (AI) technologies has led to the emergence of various intelligent applications, including DeepSeek. With over 10 million downloads and thousands of user reviews on the Google Play Store, DeepSeek has garnered significant public attention. These user reviews provide valuable insights into perceptions of the app’s performance, highlighting the need for systematic analysis to support service improvement. Furthermore, controversies and restrictions surrounding DeepSeek in certain countries due to data security and privacy concerns underscore the importance of understanding public sentiment more deeply. This study employs the Support Vector Machine (SVM) algorithm to classify the sentiment of 1,000 user reviews obtained via web scraping. The dataset was processed through text preprocessing stages, including case folding, cleaning, normalization, tokenization, stopword removal, and stemming. Feature extraction was performed using TF-IDF, and classification utilized the One-vs-All approach. Evaluation results demonstrate that the SVM model successfully categorized sentiments into positive, neutral, and negative classes with an accuracy of 80%. The highest F1-scores were achieved in the positive (0.84), negative (0.77), and neutral (0.75) categories. The analysis revealed 38.73% positive, 36.98% neutral, and 24.29% negative reviews. These findings confirm the effectiveness of SVM for sentiment analysis of Indonesian texts and its potential in informing AI-based product development strategies.
Lalupanda et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: