本研究利用支持向量機(SVM)來預測氣象資訊對美髮沙龍洗療護產品銷量預測分析。透過已知的氣候中央氣象局開放相關資料,包括最低溫度、最高溫度、平均溫度、平均雨量、風速最大十分鐘、最大瞬間風、氣壓、平均濕度、降雨日數、日照時數等資料,利用LibSVM將氣象資訊對美髮沙龍洗療護產品銷量進行預測分析。本研究蒐集台中市美髮業者及髮廊,六年內的氣象資訊及洗療產品的年度月實售數量(民國100年1月至民國106年)為實驗數據,將利用LibSVM發展預測模型。研究結果顯示支持向量機可以準確地預測出氣象資訊對美髮沙龍洗療護產品銷量預測分析,分類結果具高度的準確性。實驗結果能協助洗療護產品相關經銷公司在設計行銷活動或進貨庫存時,提供有效相關資訊,讓店家能對存貨的掌握有更多的參考依據,進而降低庫存成本,增加營收利益。
In this study, we adopt support vector machine (SVM) to predict the hair salon wash care product sales volume according to the weather information. With climate-related open data of the Central Weather Bureau including the minimum temperature, maximum temperature, average temperature, average rainfall, wind speed, maximum ten minutes, the maximum instantaneous wind, pressure, humidity, average rainfall for several days and sunshine hours, those weather information are used as the attributes of LibSVM to run the analysis of sales volume prediction. The number of monthly sales volume of the hair salon wash care products from Jan. of 2011 to April of 2016 are collected from Taichung hairdressing salon industry by combining the weather information from Central Weather Burea as the experimental data for the prediction model. The results show that support vector machines can accurately predict the weather information for the sales volume of the hair salon wash care products. The results can help the company of the hair salon wash care products to plan good marketing campaigns and reduce inventory costs.