以M-GSPRT法應用於家庭消費產品需求預測之研究∼以休旅車為例

Abstract

過去消費行為研究常以個人決策的觀點出發，然而某些家庭消費產品（如汽車、家俱）其消費決策往往是由家庭群體決策行為所產生的，故以群體決策的觀念導入家庭消費產品的需求預測是合理且必須的，而這樣的需求預測方法是之前未被提出的。因此本研究以M-GSPRT法為主並結合貝斯估計學習法提出一以家庭群體決策為基礎的家庭消費產品需求預測方法，並以休旅車做為案例研究產品。M-GSPRT法以序列式輸入樣本的特色，將家庭各決策成員的產品需求意見依決策力高低依序考慮，意義就如同決策力較高的成員擁有先決的權力，與家庭決策的實際情形是接近的。休旅車案例研究中發現，家庭在休旅車的消費決策中普遍存在家庭群體決策的性質；而休旅車產品的需求預測結果方面，『操控性』、『安全性』、『乘坐舒適性』是在整體及各區隔家庭下皆受到高度需求的，而『越野能力』則未在任何一個家庭區隔中呈現高度需求。

關鍵字：群體決策、M-GSPRT法、休旅車

Despite the fact that in previous literature, consumer behavior methodologies are oriented from the viewpoint of individual decision, some consuming maneuvers (such as cars and furniture) are determined by household group decisions. Hence, it is necessary to employ the concept of group decision making to forecast household-based consuming behavior. Accordingly, aiming at the example of Recreational Vehicle (RV) products, this study investigates a household-based consuming behavior forecasting approach M-GSPRT techniques coupled with Bayesian Estimation & Learning techniques. The feature of the proposed method is that it inputs the sample sequentially, and considers each household member’s demand opinions sequentially depending on the member’s decision power. This means that the member with more decision power has relatively higher decision priority. Our numerical results show the applicability of the proposed method for household’s consuming behavior forecasting. In addition, the results of demand forecasting show that the product features, including “Handling”, “Safety”, and “Riding comfort”, are all highly demanded for whole family; the function of “Off-road ability” is least important.

Key words：Group decision , M-GSPRT（Modified Generalized Sequential Probability Ratio Test）, Recreational Vehicle（RV）