In the era of data, machine learning techniques are increasingly applied across various fields, particularly showing great potential in handling classification problems. This study combines machine learning methods with sensory analysis techniques, using Correspondence Analysis (CA), Decision Trees, and Random Forests. It also employs Shapiro-Wilk test, Friedman test, Pairwise Wilcoxon test, Anderson-Darling test, and Kruskal-Wallis test to explore classification methods and results for two different data sets.The first data set comes from a headphone listening experiment, aimed at studying the classification effectiveness of decision tree models on headphone attributes. The results indicate that, after pruning, the four headphones can be distinctly classified into different categories, consistent with the results of Correspondence Analysis and objective SPL measurement curves. After further parameter optimization, the decision tree model performed well in terms of accuracy and Kappa value under specific settings, but the Random Forest model was superior in overall accuracy and stability, providing clearer classification of the headphones.The second data set comes from a business district survey experiment, aimed at analyzing the differences in opinions between men and women regarding the Fengjia business district. The decision tree model clearly categorized men's and women's views on parking convenience. The results showed that men generally believed that parking needed improvement, while women were relatively satisfied. For classifications of significant and highly significant views, both the decision tree and random forest models were compared and analyzed. Although the decision tree model performed well under specific settings, the random forest model outperformed in terms of accuracy and Kappa value, and provided clearer results for gender classification.Overall, this study demonstrates, through a comparison of decision tree and random forest models, the superiority of random forests in the classification of headphone sound quality and opinions about the business district. These findings offer valuable insights for further optimizing headphone design and understanding consumer behavior in business districts.
第一章 緒論 1
第二章 文獻探討 5
第一節 感官分析與商圈調查研究 5
第二節 機器學習分類方法 7
第三章 研究方法介紹 11
第一節 決策樹 11
第二節 隨機森林 26
第三節 對應分析 29
第四章 實證結果與分析 39
第一節 感官測試資料分析 39
第二節 逢甲商圈調查資料分析 60
第五章 結論與建議 71
參考文獻 73
附錄 77
附錄一 聽覺感官問卷設計 77
附錄二 逢甲商圈調查問卷 81

中文文獻：
楊亨利、林青峰. (2020)，『針對情感商品的推薦機制－以流行音樂為 例』，中華民國資訊管理學報，第二十七巻，第二期，頁 175-204。
蔡儒儀. (2022)。消費者對臺灣市售高粱酒的感官特徵感受及其接受性之研究。﹝碩士論文。國立金門大學﹞臺灣博碩士論文知識加值系統。
黃建超. (2021). 遊客對台南市安平老街商圈重視度與滿意度比較之研究. 城市學學刊, 11(1), 205-240.
李筑軒 .(2024). 形象商圈意象, 服務品質之影響分析-以臺中逢甲商圈為例. 休閒研究, 14(1), 132-152.
黃柏凱, 洪逸安, & 朱啓銘. (2022). 特殊情況下飲食空間的認知調查-以中原商圈為例. 2022 室內設計教育國際論壇論文集, b11-03.
英文文獻：
Arrieta Sagredo, I., Moulin, S., & Bech, S. (2019). Sensory profiling of high-end loudspeakers using rapid methods - Part 4: Flash Profile with expert assessors. In Proceedings of the 146th Convention (Convention Paper 10134). Presented at the 146th Convention, March 20–23, 2019, Dublin, Ireland. Audio Engineering Society.
Aldi, F., & Rahma, A. A. (2019). University Student Satisfaction Analysis on Academic Services by Using Decision Tree C4.5 Algorithm (Case Study: Universitas Putra Indonesia "YPTK" Padang). Journal of Physics: Conference Series, 1339, 012051. doi:10.1088/1742-6596/1339/1/012051
Cao, Z., Chen, T., & Cao, Y. (2021). Effect of Occupational Health and Safety Training for Chinese Construction Workers Based on the CHAID Decision Tree.　Frontiers　in　Public　Health,　9,　623441.　doi:　10.3389/fpubh.2021.623441.
Duan, L., He, J., Li, M., Dai, J., Zhou, Y., Lai, F., & Zhu, G. (2021). Based on a Decision Tree Model for Exploring the Risk Factors of Smartphone Addiction Among Children and Adolescents in China During the COVID-19Pandemic.Frontiers　in　Psychiatry,　12,　652356.　doi:　10.3389/fpsyt.2021.652356.
Di Cairano, M., Condelli, N., Galgano, F., & Caruso, M. C. (2022). Experimental gluten-free biscuits with underexploited flours versus commercial products: Preference pattern and sensory characterisation by Check All That Apply Questionnaire. International Journal of Food Science and Technology, 57, 1936–1944.
Greenacre, M. (2016). Correspondence Analysis in Practice. CRC Press.
Giacalone, D., Nitkiewicz, M., Moulin, S., Boðason, T., Laugesen, J. L., & Bech, S. (2017). Sensory profiling of high-end loudspeakers using rapid methods - Part 2: Projective mapping with expert and naïve assessors. In Proceedings of the 142nd Convention (Convention Paper 9775, pp. 1-5). Presented at the 142nd Convention, May 20–23, 2017, Berlin, Germany. Audio Engineering Society.
Guzzetti, A., Crespi, R., & Belvedere, V. (2023). Phygital luxury experiences: A correspondence analysis on retail technologies. International Journal of Consumer　Studies.　Advance　online　publication.
Available online:https://doi.org/10.1111/ijcs.13008
Hicks, L., Moulin, S., & Bech, S. (2018). Sensory Profiling of High-End Loudspeakers Using Rapid Methods–Part 3: Check-All-That-Apply with Naïve Assessors. Journal of the Audio Engineering Society, 66(5), 329–342.
Kroese, D. P., Botev, Z. I., Taimre, T., & Vaisman, R. (2022, May 8). Data Science and Machine Learning: Mathematical and Statistical Methods. Chapman and Hall/CRC.
Lee, S., Kwak, H. S., Kim, S. S., & Lee, Y. (2021). Combination of the Check-All-That-Apply (CATA) Method and Just-About-Right (JAR) Scale to Evaluate Korean Traditional Rice Wine (Yakju). Foods, 10, 1895.
Moulin, S., Bech, S., & Stegenborg-Andersen, T. (2016). Sensory profiling of high-end loudspeakers using rapid methods - Part 1: Baseline experiment using headphone reproduction. In Proceedings of the Conference on Headphone Technology (pp. 1-5). Presented at the Conference on Headphone Technology, August 24–26, 2016, Aalborg, Denmark. Audio Engineering Society.
Matzavela, V., & Alepis, E. (2021). Decision tree learning through a Predictive Model for Student Academic Performance in Intelligent M-Learning environments. Computers and Education: Artificial Intelligence, 2, 100035.
Madaan, M., Kumar, A., Keshri, C., Jain, R., & Nagrath, P. (2021). Loan default prediction using decision trees and random forest: A comparative study. In IOP conference series: materials science and engineering (Vol. 1022, No. 1, p. 012042). IOP Publishing.
Matejková, E., & Matušek, V. (2022). The use of correspondence analysis in exploring consumer purchasing behavior. Mathematics Education Research and　Applications,　8(2),　86-98.
Available online:https://doi.org/10.15414/meraa.2022.08.02.86-98
Nomura, Y., et al. (2021). Structure and Validity of Questionnaire for Oral Frail Screening.Healthcare, 9(1), 45.
Available online:https://doi.org/10.3390/healthcare9010045
Pasaribu, U. S., Lestari, K. E., Indratno, S. W., & Garmini, H. (2020). Toward Enhancement of Higher Education Quality by Alumni Tracking Using Correspondence Analysis. International Journal of Innovation, Creativity and Change, 13(10), 464.
Pathy, G. S., Ramanathan, H. N., & Iyer, E. (2021). Measurement of buying roles in family decision-making process for gold jewellery using correspondence analysis. PJAEE, 18(6).
Roseline, S. A., Geetha, S., Kadry, S., & Nam, Y. (2020). Intelligent vision-based malware detection and classification using deep random forest paradigm. IEEE Access, 8, 206303-206324.
Siahaan, H., et al. (2019). Application of Classification Method C4.5 on Selection of Exemplary Teachers. Journal of Physics: Conference Series, 1235, 012005. doi:10.1088/1742-6596/1235/1/012005.
Salazar-Concha, C., & Ramírez-Correa, P. (2021). Predicting the Intention to Donate Blood among Blood Donors Using a Decision Tree Algorithm. Symmetry, 13, 1460. Available online: https://doi.org/10.3390/sym13081460
Salamaga, M. (2021). The use of correspondence analysis in the study of foreign divestment in the Visegrad Countries during the coronavirus crisis. Econometrics. Ekonometria. Advances in Applied Data Analysis, 25(2). Available online:https://doi.org/10.15611/eada.2021.2.02
Tiggaa, N. P., & Garg, S. (2020). Prediction of Type 2 Diabetes using Machine Learning Classification Methods. Procedia Computer Science, 167, 706–716.
Wiwiek Katrina, et al. (2019). C.45 Classification Rules Model for Determining Students' Level of Understanding of the Subject. Journal of Physics: Conference Series, 1255, 012005. doi:10.1088/1742-6596/1255/1/012005
Widyastuti, M., et al. (2019). Classification Model C.45 on Determining the Quality of Customer Service in Bank BTN Pematangsiantar Branch. Journal of Physics: Conference Series, 1255, 012002. doi:10.1088/1742-6596/1255/1/012002.
Wang, X., Zhai, M., Ren, Z., Ren, H., Li, M., Quan, D., ... & Qiu, L. (2021). Exploratory study on classification of diabetes mellitus through a combined Random Forest Classifier. BMC medical informatics and decision making, 21, 1-14.
Zańko, A., Milewska, K., & Milewski, R. (2020). Correspondence Analysis in the Assessment of the Influence of Lifestyle on Infertility of Various Origins. Studies　in　Logic,　Grammar　and　Rhetoric,　64(77).Available online:https://doi.org/10.2478/slgr-2020-0038
Giacalone, D., Nitkiewicz, M., Moulin, S., Boˇgason, T., Laugesen, J. L., and Bech, S., “Sensory profiling of high-end loudspeakers using rapid methods - Part 2,” 142nd Audio Engineering Society International Convention 2017, Aes 2017, 2017.
Hicks, L., Moulin, S., and Bech, S., “Sensory profiling of high-end loudspeakers using rapid methods - Part 3: Check-all-that-apply with naïve assessors,” Journal of the Audio Engineering Society, 66(5), pp. 329–342, 2018, ISSN 1549-4950, doi:10.17743/jaes.2018.00