Bank BTN is the leading provider of mortgage (KPR) services in Indonesia, yet customer complaints on social media remain frequent. According to the OJK 2023 Annual Report, the banking sector recorded 10,848 consumer complaints, an increase from 7,425 complaints in 2022. This trend highlights the need for a systematic analysis of public opinion on KPR services. This study analyzes public sentiment expressed on Bank BTN’s official Instagram account during 2023. User comments were classified into positive and negative sentiments using text mining techniques. Four machine learning algorithms were compared: K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Decision Tree (DT), and Naive Bayes (NB). A total of 1,507 comments were obtained after data cleaning, preprocessing, and filtering from 28,745 raw comments. SMOTE was applied to address the class imbalance. Decision Tree achieved the highest F1-score (0.9348) on the imbalanced test set, followed by SVM (0.9259). Trend analysis was conducted using Pearson, Spearman, and Kendall correlation tests. The correlation between daily sentiment and KPR disbursement was weak and not statistically significant (Pearson: 0.082, p = 0.187; Spearman: 0.099, p = 0.112; Kendall: 0.077, p = 0.102; N = 259). These results indicate that sentiment fluctuations on Instagram cannot be used as a direct predictor of same-day mortgage disbursement. There may be a time lag, a non-linear relationship structure, or the influence of more dominant fundamental factors.
Copyrights © 2026