| Summary: | 碩士 === 國立中正大學 === 資訊管理學系暨研究所 === 102 === In this past few years by the rapid rise of China industries, Even affect the survival of Taiwan's dynamic manufacturing sector, To keep maintain Taiwan's industrial competitiveness, stabilize Taiwan's industrial advantages, the quality, development and yield will be an important topic. Since 2008 the solar industry to play a more vicious competition in the China market is affected, causing serious injuries worldwide, making the single & poly silicon solar price to attract buyers, the cost is closer to Thin-film solar energy with low-cost advantage. Thin-film solar cell conversion efficiency of less than one drawback single & poly silicon sun was out to enlarge the review, making the thin-film technology , not been widely respected and also affect the entire market. Thin-film solar industry, in order to improve the conversion efficiency place of the grab market, using a variety of techniques and analysis of factors related to attempting to find the conversion efficiency from the manufacturing process , improving the overall manufacturing process.
In the manufacturing process the collected data available, the use of scientific methods of data mining technology, quickly find the key factor in a big data, the purpose of this study is also proved that over the years , relying on the judgment of expert method to determine the key attributes, is not the best approach. Instead, the use of dimensionality reduction technology for multiple stepwise regression analysis methods, combined with performance improvement classification methods such as two ways to experiment, to find a better prediction model from the experiment.
|
|---|
