“Our company has worked with GuanData on three different Machine Learning projects during the last year, and in each case was a pleasure to work with.  The team is technically sound and always on the forefront of new modelling techniques.  Deliverables are always on time and on brief.  Rather than simply being a vendor, GuanData is dedicated to working side by side with ABI as a trusted partner.  And most importantly, the team enjoys its work, is eager to drive results, and delivers.  When future projects arise we would be happy to work with GuanData again.”

背景介紹

行業(yè)：快消品
業(yè)務(wù)：全國有430多家KA售點，旗下商品種類超過820個，與KA商戶建立協(xié)同式供應(yīng)鏈庫存管理模式（CPFR）。

用戶痛點

在預測層面，目前更多停留在專家經(jīng)驗判斷，人工預測費時費力，且面臨預測專家流失問題。

解決方案

AI預測：對歷史數(shù)據(jù)進行正則化、log預處理的分析探索，融入地理位置、門店經(jīng)濟指數(shù)、氣溫、賽事、節(jié)假日、促銷活動等關(guān)鍵因子，基于時間序列、樹模型、深度學習進行模型融合，輸出預測結(jié)果。

觀遠價值

固化經(jīng)驗：將專家經(jīng)驗固化在系統(tǒng)里，使其成為企業(yè)長久的數(shù)據(jù)資產(chǎn)。
提升效率：大幅降低人工預測成本，減輕日常數(shù)據(jù)繁瑣的計算工作，提升3倍工作效率。
提高準度：比專家預測提升5%的精準度，大幅減少企業(yè)的生產(chǎn)、物流、倉儲等成本， 提升了對售點突發(fā)事件（比如促銷、節(jié)日、大型活動等）的應(yīng)對能力。