智慧語詞情緒識別之自我檢測系統

Self-detection System of Intelligent Vocabulary Emotion Recognition

陳佳鈞(Chia-Chun Chen)

指導教授 ： 徐豐明

國立臺中科技大學/設計學院/多媒體設計系碩士班/碩士(2021年)

摘要

現代人生活壓力越來越大，人們在家庭、工作、人際相處日常生活中日益緊繃，導致現今社會因壓力過大引發精神疾病而自殺的事件頻傳，然而其中有部分原因是拜科技進步所賜。科技進步可為生活帶來許多便利性，同時也為人們帶來許多不可避免的困擾。因此本研究欲探討這些長期處在壓力影響所導致的精神疾病是否能以人工智慧系統提供初步協助，幫助使用者識別自己的情緒，預防精神疾病的產生。

人工智慧是近幾年的研究趨勢，其中不乏許多聊天機器人的發展應用，因其在未來的運用市場商機無限，不論是輔助調查使用者喜好還是協助資訊傳達都具有非常顯著的效果，但目前國內研究卻鮮少有人研究具判定情緒功能的聊天機器人，因此本研究使用機器學習結合自然語言處理和Jieba的技術進行語音情緒識別。

該系統在進行使用者測試時具有一定的準確率與應用性，甚至高於系統本身的效能評估。此系統提供大眾族群一個自身心理健康狀態初步檢測之管道，當使用者做檢測時，就可以初步評估自身心狀態，並根據結果分析得知目前身心狀況是否健康與後續處理的方式，此結果可作為使用者是否需要進一步評估心理健康狀態及需要專業醫療協助的初步指標，也可作為輔助提供諮商師關於使用者的相關資訊，幫助諮商師在處理個案時能更快瞭解個案目前狀況。

關鍵字

機器學習 ； 語音情緒識別 ； 自然語言處理 ； Jieba ； 了解自我情緒

並列摘要

The life pressure of modern people is increasing, and people's daily life in family, work and interpersonal relationships is becoming more and more tense. Due to excessive stress and mental illness, this has led to frequent suicides in today's society. However, part of the reason is due to advances in science and technology. The advancement of science and technology can bring a lot of convenience to life, but it also brings many unavoidable troubles to people. Therefore, this study aims to explore whether these mental illnesses caused by long-term stress can use artificial intelligence systems to provide initial help to help users understand their emotions and prevent mental illness.

Artificial intelligence is a research trend in recent years. There is not only of chatbot development and applications. Since there are unlimited market opportunities in the future, it plays a very important role in assisting in investigating user preferences or assisting in information transmission. However, few people have studied chatbots that can determine emotions. Therefore, this research combines machine learning with natural language processing and Jieba's speech emotion recognition technology.

The system has a certain accuracy and applicability in the user test process, even higher than the performance evaluation of the system itself. The system provides a preliminary test of the mental health of the general population. When the user conducts the test, he can initially assess his mental state and analyze the results to understand whether the current physical and mental condition is healthy and how to deal with it. As a preliminary indicator of whether users need to further evaluate their mental health status and whether they need professional medical help, it can also be used as an auxiliary tool to provide counselors with relevant information about users to help counselors understand the patient’s current condition. When handling cases, let case processing speed is faster.

並列關鍵字

Machine Learning ； Speech Emotion Recognition ； Natural Language Processing ； Jieba ； Understanding Self Emotions

參考文獻

[1]M. Mori, "The uncanny valley, " Energy, vol. 7, no. 4, pp. 33–35, 1970.

[2]Adang Suwandi Ahmad and Arwin Datumaya Wahyudi Sumari, "Cognitive Artificial Intelligence: Brain-Inspired Intelligent Computation in Artificial Intelligence, " Computing Conference 2017 pp. 18-20, July, 2017.

[3]Xingbao Gao and Li-Zhi Liao, "A New One-Layer Neural Network for Linear and Quadratic Programming, " Ieee Transactions On Neural Networks, vol. 21, no. 6, pp. 918-929, June, 2010.

[4]Xuefeng Zhang, Wenying Wang, Xuanxuan Zheng, Yongxue Ma, Yao Wei, Ming Li and Yu Zhang, "A Clutter Suppression Method Based on SOM-SMOTE Random Forest," 2019 IEEE Radar Conference (RadarConf), Boston, MA, USA, pp. 1-4, 2019.

[5]Y. Bengio, P. Simard and P. Frasconi, "Learning long-term dependencies with gradient descent is difficult, " in IEEE Transactions on Neural Networks, vol. 5, no. 2, pp. 157-166, March 1994.