近年來，隨著機器學習模型的日益複雜化，各領域對於模型解釋性的需求不斷增加，尤其是在需要理解模型決策的關鍵領域。本研究專注於開發一個框架，用於從高維度文本數據中提取因果洞見。透過運用先進的因果推論和反事實生成技術，該方法能夠識別並量化關鍵文本特徵對預測結果的影響，提供每個特徵對模型決策影響的清晰指標。不同於傳統僅將文本視為未分化輸入的模型，這種方法允許更細緻的分析，使使用者能夠理解特定詞語或短語如何驅動模型的預測。本研究旨在縮小文本類AI系統中準確性與解釋性之間的差距，促進可解釋AI的發展，並為情感分析、風險評估和客戶回饋解讀等應用領域提供實用的見解。

In recent years, the increasing complexity of machine learning models has driven the need for greater interpretability, particularly in fields where understanding model decisions is crucial. This study focuses on developing a framework for extracting causal insights from high-dimensional textual data. By leveraging advanced techniques in causal inference and counterfactual generation, the proposed approach identifies and quantifies the impact of key textual features on prediction outcomes, providing a clear indicator of each feature’s influ-ence on model decisions. Unlike traditional models that treat text as an undifferentiated input, this method allows for a more granular analysis, enabling users to understand how specific words or phrases drive model predictions. This research aims to bridge the gap between accuracy and interpretability in text-based AI systems, contributing to the advancement of explainable AI and offering practical insights for applications in fields like sentiment analy-sis, risk assessment, and customer feedback interpretation.

Aizawa, A. (2003). An information-theoretic perspective of tf–idf measures. Information Processing & Management, 39(1), 45-65.
Bhattacharjee, A., Moraffah, R., Garland, J., & Liu, H. (2024). Zero-shot LLM-guided Counterfactual Generation for Text. arXiv preprint arXiv:2405.04793.
Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent dirichlet allocation. Journal of machine Learning research, 3(Jan), 993-1022.
Chipman, H. A., George, E. I., & McCulloch, R. E. (2010). BART: Bayesian additive regression trees.
Cui, Y., Pu, H., Shi, X., Miao, W., & Tchetgen Tchetgen, E. (2024). Semiparametric proximal causal inference. Journal of the American Statistical Association, 119(546), 1348-1359.
Devlin, J. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.
Dwivedi, R., Dave, D., Naik, H., Singhal, S., Omer, R., Patel, P., . . . Morgan, G. (2023). Explainable AI (XAI): Core ideas, techniques, and solutions. ACM Computing Surveys, 55(9), 1-33.
Feder, A., Keith, K. A., Manzoor, E., Pryzant, R., Sridhar, D., Wood-Doughty, Z., . . . Roberts, M. E. (2022). Causal inference in natural language processing: Estimation, prediction, interpretation and beyond. Transactions of the Association for Computational Linguistics, 10, 1138-1158.
Grootendorst, M. (2022). BERTopic: Neural topic modeling with a class-based TF-IDF procedure. arXiv preprint arXiv:2203.05794.
Imbens, G. W., & Rubin, D. B. (2015). Causal inference in statistics, social, and biomedical sciences: Cambridge university press.
Kingma, D. P. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.
Madaan, N., Padhi, I., Panwar, N., & Saha, D. (2021). Generate your counterfactuals: Towards controlled counterfactual generation for text. Paper presented at the Proceedings of the AAAI Conference on Artificial Intelligence.
McInnes, L., Healy, J., & Melville, J. (2018). Umap: Uniform manifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426.
Meyboom, R. H., Hekster, Y. A., Egberts, A. C., Gribnau, F. W., & Edwards, I. R. (1997). Causal or casual? The role of causality assessment in pharmacovigilance. Drug safety, 17, 374-389.
Morgan, S. (2015). Counterfactuals and causal inference: Cambridge University Press.
Mothilal, R. K., Sharma, A., & Tan, C. (2019). Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations.(5 2019). In.
Papineni, K., Roukos, S., Ward, T., & Zhu, W.-J. (2002). Bleu: a method for automatic evaluation of machine translation. Paper presented at the Proceedings of the 40th annual meeting of the Association for Computational Linguistics.
Pearl, J. (2010). Causal inference. Causality: objectives and assessment, 39-58.
Rahman, M. F., Liu, W., Suhaim, S. B., Thirumuruganathan, S., Zhang, N., & Das, G. (2016). Hdbscan: Density based clustering over location based services. arXiv preprint arXiv:1602.03730.
Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41-55.
Rubin, D. B. (1974). Estimating causal effects of treatments in randomized and nonrandomized studies. Journal of educational Psychology, 66(5), 688.
Shi, C., Blei, D., & Veitch, V. (2019). Adapting neural networks for the estimation of treatment effects. Advances in neural information processing systems, 32.
Sun, C., Qiu, X., Xu, Y., & Huang, X. (2019). How to fine-tune bert for text classification? Paper presented at the Chinese computational linguistics: 18th China national conference, CCL 2019, Kunming, China, October 18–20, 2019, proceedings 18.
Vaswani, A. (2017). Attention is all you need. Advances in neural information processing systems.
Wachter, S., Mittelstadt, B., & Russell, C. (2017). Counterfactual explanations without opening the black box: Automated decisions and the GDPR. Harv. JL & Tech., 31, 841.
Zhao, Q., & Hastie, T. (2021). Causal interpretations of black-box models. Journal of Business & Economic Statistics, 39(1), 272-281.