這篇文章分析了原料投入對於廢棄物產出量的影響，並將焦點關注積體電路產 業中的氟化鈣污泥這項重要的廢棄物。文章中使用 LASSO (Tibshirani 1996) 作 為主要的分析工具。LASSO 能讓我們辨認出對廢棄物產出有重大影響的六種 原料，分別是氟化氫、光阻、合成氨、磊晶矽晶圓、濃硫酸以及異丙醇。而且， LASSO 也能得到比最小平方法 (ordinary least squares, OLS) 更精準的預測表 現。另外，統計上的顯著性也是做研究中普遍關心的重點。本文中使用了由 Lee et al. (2016) 提出的統計推論方法，這個方法適用於在用 LASSO 選出模型後做 統計推論。由 LASSO 選出的六種原料皆是顯著的。更重要的是文章中的分析架 構能更廣泛的被用於其他方面，像是積體電路產業中其他的廢棄物或者是其他 製造業的廢棄物。

This paper studies how materials input has impact on output of calcium fluoride (CaF2) sludge which is an important type of waste in IC industry. By applying LASSO model (Tibshirani 1996), we identify 6 materials essential to calcium flu- oride sludge output including hydrogen fluoride, photoresist, synthetic ammonia, epitaxial wafer, concentrated sulfuric acid and isopropanol, and make a more ac- curate prediction about waste generation than the ordinary least squares (OLS) model does. In order to obtain statistical significance, this work perform a post- selection inference (Lee et al. 2016) which is designed especially for LASSO. We find that all materials selected by LASSO are all statistically significant. Further- more, the procedures implemented in this paper can be applied under a wide range of circumstances such as other types of wastes in other manufacturing industries.

Berk, R., Brown, L., Buja, A., Zhang, K., Zhao, L. et al. (2013). Valid post- selection inference. The Annals of Statistics, 41(2), 802–837.

Casella, G. & Berger, R. L. (2002). Statistical inference. Duxbury Pacific Grove, CA.

Efron, B. & Hastie, T. (2016). Computer age statistical inference. Cambridge Uni- versity Press.

Friedman, J., Hastie, T., & Tibshirani, R. (2010). Regularization paths for gener- alized linear models via coordinate descent. Journal of Statistical Software, 33(1), 1–22. Retrieved from http://www.jstatsoft.org/v33/i01/

Hastie, T., Tibshirani, R., & Friedman, J. (2009). The elements of statistical learn- ing: Data mining, inference, and prediction, springer series in statistics. Springer New York.

Hoerl, A. E. & Kennard, R. W. (1970). Ridge regression: Biased estimation for nonorthogonal problems. Technometrics, 12(1), 55–67.

James, G., Witten, D., Hastie, T., & Tibshirani, R. (2013). An introduction to statistical learning. Springer.

Lee, J. D., Sun, D. L., Sun, Y., Taylor, J. E. et al. (2016). Exact post-selection inference, with application to the lasso. The Annals of Statistics, 44(3), 907– 927.

Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B (Methodological), 58(1), 267–288.

Tibshirani, R., Wainwright, M., & Hastie, T. (2015). Statistical learning with spar- sity: The lasso and generalizations. Chapman and Hall/CRC.

Tibshirani, R. J. et al. (2013). The lasso problem and uniqueness. Electronic Jour- nal of Statistics, 7, 1456–1490.38

Tibshirani, R., Tibshirani, R., Taylor, J., Loftus, J., & Reid, S. (2017). Selectiveinference: Tools for post-selection inference. R package version 1.2.4. Retrieved from https://CRAN.R-project.org/package=selectiveInference

Tseng, P. et al. (1988). Coordinate ascent for maximizing nondifferentiable concave functions.

Tseng, P. (2001). Convergence of a block coordinate descent method for nondif- ferentiable minimization. Journal of optimization theory and applications, 109(3), 475–494.

Xiao, H. (2012). Introduction to semiconductor manufacturing technology.

財團法人工業技術研究院 (ITRI). (2015). 永續物料管理制度推動及系統功能提升專案工作計畫報告書.