以模糊模型探勘憂鬱症患者日常對話之關聯性

Using Fuzzy FP-tree Model to Discover Associations Among Melancholiac Patient’s Daily Text Messages

邱鴻文(Hong-Wen Chiu)

指導教授 ： 黃有評

國立臺北科技大學/電資學院/電機工程系研究所/碩士(2010年)

https://doi.org/10.6841/NTUT.2010.00629

摘要

由於現在社會的工作壓力越來越大，伴隨著經濟、家庭等民生問題，許多人開始有憂鬱的傾向和情緒上的問題，嚴重的話最後可能會演變成憂鬱症或者其他情緒上的缺陷。但是一般人常常容易忽略這些症狀或者甚至已經有憂鬱症自己卻沒發覺。為了幫助這些情緒異常的病患能夠與其他人維持正常的溝通，若能將情緒異常病患日常生活所說的話語記錄下來，經由後端資料庫內容分析病患的說話意涵，提供給病患一套完整的情緒分析結果，將可改善病患的人際關係並且降低因為情緒不佳而發生不幸的可能性。本篇論文將結合以斷字系統、模糊理論和資料探勘關聯法則來分析病患情緒，首先以中央研究院斷字系統對記錄下來的病患話語進行斷字處理，並利用本研究所提出的權重計算方式篩選出關鍵字，將所得到的關鍵字進行關聯法則的探勘，找出病患所說的字詞之間的關聯性，或者在說了哪些字詞之後會產生負面情緒，藉由這些關聯性可進一步了解到病患本身所面臨的問題，醫生即可參考這些問題，對病患做心理輔導與建設。最後，本研究將以滑動式窗的方式對病患分析的結果，製作病患的情緒起伏表，以觀察病患長期的情緒變化，進而預測未來病患可能的情緒反應。

關鍵字

資料探勘 ； 關聯法則 ； 模糊理論 ； 斷詞

並列摘要

With the constant stress from work load and daily life such as economic and family problems people may show symptoms of melancholia. Even worse, it may eventually turn into depression. However, most people are reluctant to describe it or may not know that they already have it. To help these patients maintain communication normally with other persons, the patients’ daily life text messages are recorded for further analysis. The analytical results would be useful for patients to improve their own social relationship and to reduce the unfortunate occurrence of poor emotional possibilities. In this study a novel system is proposed to analyze patients’ emotion using word segmentation system and fuzzy association rules. The word segmentation system is provided by Academia Sinica, Taiwan for users to input text messages. Critical keywords are selected by the proposed weighted keyword strategy. A fuzzy association model is used to discover frequent keywords mentioned in the speech files and to find out the relation among them. Based on these extracted associations the proposed system can help figure out what the problems a patient has faced. Then, the doctors can give patients some psychological counseling and advice based on these results. Finally, this study presents the analytical results to patients at a fixed interval. The patients’ emotional tables are produced to observe their long-term variations in emotion and to predict any possible emotion reactions in the future.

並列關鍵字

data mining ； association rules ； fuzzy set theory ； word segmentation

參考文獻

[1] A. J. Mitchell and M. Shiri-Feshki, “Temporal Trends in the Long Term Risk of Progression of Mild Cognitive Impairment: a Pooled Analysis,” Journal of Neurology, Neurosurgery and Psychiatry, vol. 79, no. 12, June 2008, pp.1386-1391.

[2] R. S. Wilson, N. T. Aggarwal, L. L. Barnes, J. L. Bienias, C. F. Mendes de Leon, and D. A. Evans, “Biracial Population Study of Mortality in Mild Cognitive Impairment and Alzheimer Disease,” Archives of Neurology, vol. 66, no. 6, June 2009, pp.767-772.

[4] D. Perera, J. Kay, I. W. Koprinska and K. Yacef, “Clustering and Sequential Pattern Mining of Online Collaborative Learning Data,” IEEE Trans. on Knowledge and Data Engineering, vol. 21, no. 6, June 2009, pp.759-772.

[10] A. Wong and G. Li, “Simultaneous Pattern and Data Clustering for Pattern Cluster Analysis,” IEEE Trans. on Knowledge Analysis and Data Engineering, vol. 20, no. 7, July 2008, pp.911-923.

[11] Y. Huang and L. Zhang, “A Framework for Mining Sequential Patterns from Spatio-Temporal Event Data Sets,” IEEE Trans. on Knowledge Analysis and Data Engineering, vol. 20, no. 4, April 2008, pp.433-448.