社群大數據－重建消費者信心指標於車市需求預測

Abstract

隨著社群網路的成熟，現代人能透過社群平台輕易地傳達自己的意見，而這些社群數據經過整理、分析後變成了「網路聲量」－一個能夠反應網民對某一主題態度的指標。本研究利用OpView社群口碑資料庫蒐集之網路聲量，並參考中央大學消費者信心指數的調查方式重建「消費者信心聲量」，以期透過社群大數據建構一個能更貼切反應民眾對當前經濟情勢看法的信心指標。 本研究並以重建之消費者信心聲量探討該指標與汽車掛牌數之關聯，比較消費者信心指數與消費者信心聲量兩者對於耐久財購買的預測效力。研究結果顯示消費者信心聲量對於汽車需求具有領先10個月的預測能力，同時也發現消費者信心聲量對於未來耐久財的購買的預測能力優於傳統消費者信心指數，足以表示本研究重建之消費者信心聲量除了能提供執政者作為民眾對未來經濟情勢看法之參考外，也能作為企業進行耐久財產品銷售預測的參數。

As social networking sites become a widely accepted part of many people’s lives, people could convey their opinions easier than before. This allows us to capture unprecedented data, Internet sentiments, which can be further explored and analyzed. This research intends to use Internet sentiments to rebuild consumer confidence index, in order to capture true thoughts of consumer on current and future economic conditions. By referring the survey method of Consumer Confidence Index conducted by National Central University, we constructed a new index, Consumer Confidence sentiment, by using social big data from Opview. In this thesis, we rebuilt Consumer Confidence Index by Internet sentiments and analyzed the correlation between the index and sales of durable goods, in here we use vehicle licensing as the proxy variable. The findings indicate that: (1) Consumer Confidence Sentiments can well predict the demand of vehicles ten months beforehand. (2) Consumer Confidence Sentiments generally out-perform Consumer Confidence Index in predicting vehicles demand. (3) Consumer Confidence Sentiments can serve as leading indicator or reference for government or corporate.