工作內容： ● 利用機器和深度學習改善工廠生產 ● 協助智慧製造專案發展與工廠並行發展 ● 處級數位轉型課程規劃 ● 內部數位轉型課程講師 專案成果： 1.冰水系統節能最佳化 ● 內容：結合空調原理和大數據，建立機器學習模型，尋找最佳節能參數。 ● 效益：節省電費(377萬元/年)及碳排放。 2.放流水加藥最佳化 ● 內容：考慮水質和流量，調整加藥量(避免浪費)，穩定pH值。 ● 效益：減少藥費(131萬元/年)。 3.鼓風機節能最佳化 ● 內容：考慮水質和流量，調節鼓風機頻率，且滿足生物需求。 ● 效益：節省電費(58萬元/年)。 4.無塵室溫濕度控制 ● 內容：求出參數建議值讓溫濕度穩定。 ● 效益：將溫度和濕度控制在5%以內。 ● 榮譽：全AUO智慧製造楷模、智慧製造_菁英Level 3人才 Responsible for: ● Improve factory production with machine and deep learning ● Help the development of smart manufacturing projects and the parallel development of the factory (Equipment warning, energy saving and fee reduction, etc.) Project achievements: i. Chilled Water Energy Saving System ● Content: Combine air conditioning principles and big data, create machine learning model, and find the best energy-saving parameters. ● Benefit: Reduce electricity bill 3,770,000 NTD/year and carbon emission. ii. Optimization of discharge water and dosing ● Content: Adjusted dosage based on water quality and pH value. ● Benefit: Reduce drug expenses 1,310,000 NTD/year. iii. Power saving of blower ● Content: Consider water quality and flow, adjust blower frequency, achieve biological needs. ● Benefit: Reduce electricity bill 580,000 NTD/year. iv. Clean room Temperature and Humidity Management ● Content: Find the parameters suggested value let temperature and humidity stability. ● Benefit: control the temperature and humidity within 5%. ● Honor: Role model of Smart manufacturing、Level 3 of Talents

工作內容： ● 參與客戶需求分析，評估技術可行性並進行數據分析、模型建立等相關技術的規劃與開發。 ● 負責數據清洗、加工、分類等開發工作，運用數據演算法及工具對數據進行挖掘工作。 ● 負責數據模型的規劃並進行數據分析，支持應用端資料分析與預測。 專案成果： 1. 預測性維護 ● 內容：結合生產資料與故障紀錄，︀分析找出即將故障的高風險機台。 ● 效益：藉由提早發現設備故障，︀提早進行維護保養，︀減少機台當機時間。 2. 訂單預測 ● 內容：結合大數據與機器學習模型預測未來訂單數量(考慮前置時間的條件下)。 ● 效益：針對預測數量進行採購，︀避免數量不足或過多造成額外成本。 3. 新進人員久任預測 ● 內容：透過分析ERP資料(個人資料、工時、薪資、打卡)，︀預測新進人員久任機率，︀ 提供部門主管招聘建議。 ● 效益：新人離職率從59%降至20%。 Responsible for: ● Responsible for data cleaning, and use data algorithms and tools to data mining. ● Participate in customer needs analysis, evaluate technical feasibility and building model. Project achievements: i. Predictive Maintenance and Quality ● Content: Calculated the best maintenance time via XGB. ● Benefit: 86% of the machine failures can be discovered early. ii. Order Forecast ● Content: Order Forecast via Machine Learning model(Lasso、Ridge、XGB). ● Benefit: Result in correctly control 70% of orders. iii. Employee turnover prediction using ERP data ● Content: Finding Your Best Candidates via Random Forest model. ● Benefit: Expected benefit is reducing the resignation rate from 59% to 20%.

1. 招生分析與推薦 ● 內容：透過網路爬蟲取得公開資料(交叉查榜)並合併私人資料(學生基本資料&成績)進行分析，為各系所提供招生建議。 ● 效益：錄取率提高8.3%。 2. 建立共用資料庫 ● 內容：整合各項填報資料(ex: 校庫、QS大學排名)所需欄位並建立資料庫。 ● 效益：減少人員重覆填報相同資訊的次數與時間。 i. Admissions Analysis ● Content: Getting the public data via Web Scraping and merge private data for analysis, provide each department suggestions for admissions. ● Benefit: Resulted in increasing the admission rate by 8.3%. ii. Create General Database ● Content: Integrate data fields for all submitted information need. ● Benefit: Reduce the number and time of personnel filling in reports.

串聯健保資料庫檔案(個人基本資料+看病或住院紀錄+用藥紀錄) 經由數據清洗、加工、分類, 分析用藥頻率與疾病之關係。 Data mining in National Health Insurance Research Database (NHIRD) via SAS and R. Analyze and predict the relationship between medication frequency and disease.