應用進化式學習架構於人格特質分析－以雲林科大大學生為例

Personality Analysis with a Multilevel Evolutionary Learning Architecture: A Case Study of NYUST Undergraduate Students

陳重臣(Jong-Chen Chen) ； 林金霖(Chin-Lin Lin) ； 張青桃(Ching-Tao Chang)

《資管評論》 9期 (1999/12) Pp. 29-51

https://doi.org/10.6131/MISRb.199912_(9).0002

摘要

近年來，國內政治及經濟的高速發展，對個人人格特質產生不小的影響。傳統的研究是從心理學、社會學、教育學、及臨床醫學等角度來探討人格特質。本文是從資訊科技人工智慧應用的角度，對人格特質資料作歸納整理及分析。為了搜集實際人格特質資料，本研究首先設計一份問卷評量表，對本校大一及大二134名學生作問卷調查，然後，利用一個三層式系統架構，並透過進化式演算法（或稱為基因演算法），以探討及分析該族群的人格特質。實驗結果顯示，系統可以由受測族群的資料中，整理出代表該族群的人格特徵。在問卷評量項目的適切性方面，系統顯現相當良好的自主性學習能力，依資料型態的不同，決定重要及不重要的評量項目。刪除不重要（或多餘）的評量項目，系統不僅可以更明顯的反應出受測族群的人格特質，而且表現對資料異動具有較高的容忍能力。

關鍵字

人格特質分析 ； 問卷評量表 ； 自主性學習 ； 進化式演算法（基因演算法）

並列摘要

The growth of politics and economy in Taiwan recently has significant impacts on human personality. Traditional approaches on the studies of personality are from the viewpoints of psychology, social science, education, and clinical medical. In this paper, the emphasis is on the application of artificial intelligence in computer science to the analysis of personality. To collect real personality data, a questionnaire was first constructed, and then conducted with 134 undergraduate students in this school. Then, a self -organizing learning system, comprising a thee-level architecture, was developed to investigate and analyze the personality of the group through evolutionary learning. Experimental results show that the system has an effective self-organizing learning capability in identifying significant an d insignificant items of the questionnaire, based on the structure of input data. Deleting insignificant items of the questionnaire allows the system to identify the personality of the group effectively. Moreover, it shows better noise tolerance capability.

並列關鍵字

personality analysis ； questionnaire ； self-organizing learning ； evolutionary learning algorithm (genetic algorithm)