最基本上,聚類分析可以用附圖嗰種方法想像:圖入面拃點當中每一粒,都喺 X 軸(表示一個變數)同 Y 軸(表示另一個變數)度有個位置,但就噉用肉眼睇都睇得出,啲點可以分做三大類(唔同色嘅點),每個聚類[歐 1]都係「個聚類入面啲點,彼此之間距離近,同時又冚唪唥都係同聚類外嘅點距離遠嘅」;聚類分析就可以想像成「同啲點油顏色,表示每點屬邊個聚類」嘅過程[2][3]。
相似度呢個概念可以話係聚類分析嘅根基。相似度係個好合符直覺嘅概念-是但搵三件物件 A B C,一個人都能夠答「A 似 B 多啲定似 C 多啲?」噉嘅問題。但喺統計學上對聚類分析嘅研究,唔同研究者用嘅相似度指標值都唔多一樣,而且呢啲唔同嘅指標,都直覺上畀人覺得係「反映得到物件之間嘅相似度」[6]:p. 2。
人工智能(AI)係電腦科學嘅一門,專研究點樣教電腦展示出好似人噉嘅有智能行為,模擬人腦做嘅資訊處理過程[57],而正如上面所述,聚類分析响理論上就好接近人腦處理語言嗰陣嘅諗嘢方式[6]:p. 1。事實係,有唔少 AI 方面嘅工作都有用到聚類分析,即係教部電腦自動噉對某啲數據做聚類分析,得出想要嘅效果[58]。
舉例說明,異常檢測[歐 41]可以話係 AI 上最直接嘅聚類分析應用之一:異常檢測指攞一大拃個案,並且搵出啲同手上嘅大多數個案都唔同嘅異常個案,好多時係因為呢啲個案有啲可疑-例如係喺醫療上探測邊啲病人身體狀況數據(血壓同心跳率等嘅變數)異常,又或者喺金融上探測邊啲交易唔尋常(例如「短時間內出現多單交易」好多時都唔尋常,可能有詐騙)... 呀噉;教電腦自動噉做異常檢測可以將睇病或者檢查金融紀錄等嘅工作自動化[59]。
Crockett, K. A., Mclean, D., Latham, A., & Alnajran, N. (2017). Cluster analysis of twitter data: A review of algorithms (PDF). In Proceedings of the 9th International Conference on Agents and Artificial Intelligence (Vol. 2, pp. 239-249). Science and Technology Publications (SCITEPRESS)/Springer Books,呢篇文綜述咗當時有嘅聚類分析演算法,攞嚟分析社交媒體上嘅留言。
Hochin, T., Hayashi, Y., Nomiya, H., & Chowdhury, M. (2018). Quasi-optimality under pseudo F statistic in clustering data. International Journal of Engineering and Technology (UAE), 7(2), 320-324,呢篇文探討用類似方差分析(ANOVA)噉嘅方法評估聚類方案-呢套方法係想計「聚類間距離」同埋「聚類內距離」之間嘅比例(pseudo-F index),用嚟評估聚類方案。
Lloyd, S. (1982). "Least squares quantization in PCM". IEEE Transactions on Information Theory. 28 (2): 129-137.
Marks, C., Jahangiri, A., & Machiani, S. G. (2019, October). Iterative DBSCAN (I-DBSCAN) to identify aggressive driving behaviors within unlabeled real-world driving data. In 2019 IEEE Intelligent Transportation Systems Conference (ITSC) (pp. 2324-2329). IEEE,用 DBSCAN 剖析啲人揸車,有提佢哋點決定啲參數。
Milligan, G. W., & Cooper, M. C. (1985). An examination of procedures for determining the number of clusters in a data set. Psychometrika, 50, 159-179.
Nkweteyim, D. L. (2018). Clustering by partitioning around medoids using distance-based similarity measures on interval-scaled variables. Nigerian Journal of Technological Development, 15(1), 1-6.
Everitt, B. S. and Bullmore, E. T. (1999). Mixture model mapping of brain activation in functional magnetic resonance images. Human Brain Mapping 7, 1-14.
Tryon, Robert C. (1939). Cluster Analysis: Correlation Profile and Orthometric (factor) Analysis for the Isolation of Unities in Mind and Personality. Edwards Brothers.
Arthur, David; Manthey, B.; Roeglin, H. (2009). "k-means has polynomial smoothed complexity". Proceedings of the 50th Symposium on Foundations of Computer Science (FOCS). arXiv:0904.1113.
Kaufman, Leonard; Rousseeuw, Peter J. (1990-03-08), "Partitioning Around Medoids (Program PAM)", Wiley Series in Probability and Statistics, Hoboken, NJ, USA: John Wiley & Sons, Inc., pp. 68-125.
Ester, Martin; Kriegel, Hans-Peter; Sander, Jörg; Xu, Xiaowei (1996). "A density-based algorithm for discovering clusters in large spatial databases with noise". In Simoudis, Evangelos; Han, Jiawei; Fayyad, Usama M. (eds.). Proceedings of the Second International Conference on Knowledge Discovery and Data Mining (KDD-96). AAAI Press. pp. 226-231.
Pfitzner, Darius; Leibbrandt, Richard; Powers, David (2009). "Characterization and evaluation of similarity measures for pairs of clusterings". Knowledge and Information Systems. Springer. 19 (3): 361-394.
Weiss, Sholom M.; Indurkhya, Nitin; Zhang, Tong; Damerau, Fred J. (2005). Text Mining: Predictive Methods for Analyzing Unstructured Information. Springer.
Petrovic, S. (2006, October). A comparison between the silhouette index and the davies-bouldin index in labelling ids clusters. In Proceedings of the 11th Nordic workshop of secure IT systems (Vol. 2006, pp. 53-64). Citeseer.
Jaiswal, D., Kaushal, V., Singh, P. K., & Biswas, A. (2020). Green market segmentation and consumer profiling: a cluster approach to an emerging consumer market. Benchmarking: An International Journal.
Farrington, D. P., Lösel, F., Boruch, R. F., Gottfredson, D. C., Mazerolle, L., Sherman, L. W., & Weisburd, D. (2019). Advancing knowledge about replication in criminology. Journal of Experimental Criminology, 15(3), 373-396.
Haining, R., Wise, S., & Ma, J. (1998). Exploratory spatial data analysis. Journal of the Royal Statistical Society: Series D (The Statistician), 47(3), 457-469.
Martin, C., Wyllie, A., & Casswell, S. (1992). Types of New Zealand drinkers and their associated alcohol-related problems. Journal of Drug Issues, 22(3), 773-796.
Tanton, J., Dodd, L. J., Woodfield, L., & Mabhala, M. (2015). Eating behaviours of British university students: A cluster analysis on a neglected issue. Advances in preventive medicine, 2015.
Crockett, K. A., Mclean, D., Latham, A., & Alnajran, N. (2017). Cluster analysis of twitter data: A review of algorithms (PDF). In Proceedings of the 9th International Conference on Agents and Artificial Intelligence (Vol. 2, pp. 239-249). Science and Technology Publications (SCITEPRESS)/Springer Books.
Williams, G. A., & Kibowski, F. (2016). Latent class analysis and latent profile analysis (PDF). Handbook of methodological approaches to community-based research: Qualitative, quantitative, and mixed methods, 143-151.