八卦類項之特徵分類研究

《易經》為中國古人的智慧，內涵博大精深且嚴謹之邏輯系統，兼具深層寓意，使用抽象的符號組合表達出世間萬種事物；在《易經》當中，八卦為《易經》之基礎，分別為乾（☰）、兌（☱）、離（☲）、震（☳）、巽（☴）、坎（☵）、艮（☶）、坤（☷）。《易經》之三大內涵為「理、象、數」，八卦類化係易經之象數中相當重要之一部分，所謂「無象不成占、無數不能斷」，占卜時可得象、數，再搭配上類化之事物，便可得占求之解。但因每個人對於易經之解釋與理解不一，且目前並沒有較為有效的方法來明確的指出事物所屬之八卦類項，因此本研究希望使用機器學習之監督式學習來解決八卦類項之類化問題，將萬物分為七個事項，為人物、情事、身體、疾病、動物、靜物、建物，並找出各事項之特徵屬性與屬性值。利用特徵屬性與屬性值建立七個事項之資料集。透過決策樹、k-近鄰演算法與支援向量機三種分類演算法，訓練出不同分類器，並比較不同分類器之效能。比較原始與重新取樣後之資料集訓練結果，在C4.5、k-近鄰演算法與支援向量機三種分類演算法中，k-近鄰演算法表現出最好的分類精確度。本研究結果提供欲對八卦類項之類化探討之研究者們提供一個較為科學性之起始；另外，對一般大眾而言，可利用研究結果之分類器得到世界上某種事物所屬之八卦類項，並利用於卜卦、風水與中醫等領域。

關鍵字

八卦；分類；監督式學習； k-近鄰演算法；支援向量機； C4.5

並列摘要

I Ching is the wisdom of Chineses ancient and contains extensice and profound logical system. There has deep implied meaning in I Ching. I Ching uses abstract symbols to present everything in the world. Bagua that created by Fu Xi is the basic of I Ching, and including Qián (☰), Duì (☱), Lí (☲), Zhèn (☳), Xùn (☴), Kǎn (☵), Gèn (☶) and Kūn (☷). The connotations of I Ching are Philosophy, Image and Number and the generalization of Bagua is the most important part of Image and Number. Image and Number will be obtained from practise divination, and collocate with things of generalization to acquire the answer of divination. Howerer, everyone has diverse explaination and different understanding of I Ching, and there did not have an effective method to specifically refer things to Bagua. Hence, this research uses supervised learning of machine learning to solve the problems of generalization of Bagua. Everything in the world have been grouped by seven subject including human, situation, body, disease, object and building, and feature attributes and property values of every subject have been found. This research utilized the feature attributes and property values to establish datasets of seven subjects. C4.5, k- nearest neighbors algorithm and support vector machine have been applied to produce different classifiers and compare the performances of classifiers. After comparing original datasets and resampling dataset, kNN presented the most efficient accuracy of classification among C4.5, kNN and SVM. This research provides an empirical origination of generalization of Bagua. The most efficient accuracy of object is closed to 90%. This result can improve the practicability of Bagua divination.

並列關鍵字

Bagua ； Classification ； Supervised Learning ； k-nearest neighbors (kNN) ； Support Vector Machine (SVM) ； C4.5

參考文獻

﹝1﹞ Alpaydın, E. (2009). Introduction to Machine Learning: The MIT Press.

﹝6﹞ Han, E.-H. S., & Karypis, G. (2000). Centroid-based document classification: Analysis and experimental results: Springer.

﹝8﹞ Kass, G. V. (1980). An exploratory technique for investigating large quantities of categorical data. Applied statistics, 119-127.

﹝10﹞ Kreßel, U. H.-G. (1999). Pairwise classification and support vector machines. Paper presented at the Advances in kernel methods.

﹝14﹞ Murthy, S. K. (1998). Automatic construction of decision trees from data: A multi-disciplinary survey. Data mining and knowledge discovery, 2(4), 345-389.

國際替代計量

八卦類項之特徵分類研究

未授權

主題瀏覽