為了設計恰當的產品、價格、溝通與通路策略以滿足消費者，行銷研究人員致力於將消費者異質性納入聯合分析中，以提升對個體參數的預測效力。在估計個體水準的參數時，研究者遭遇到最為艱難的問題就是個體水準資料的缺乏。近年來許多學者已經研究出如何同時將異質性納入行銷模式之中，並且解決個體訊息不足的問題。在這些模式之中，有限混合與層級貝氏模式被認為是優越的方法。然而，兩種方法在獲致可靠的個體參數時仍有其缺點。為了驗證結合行為模式相似的個體訊息能夠大幅地提升預測效力，本研究提出了新的層級貝氏聯合區隔模式。這個模式結合了市場區隔與層級貝氏方法，能夠解決前述兩種模式的缺點。

To design appropriate product, price, communication and channel strategies to satisfy consumers, marketing researchers endeavor to incorporate consumer heterogeneity into conjoint analyses to increase predictive validity of individual parameters. In estimating individual-level parameters, the most difficult problem facing researchers is the deficiency of individual-level data. In recent years, many scholars have already studied how to incorporate heterogeneity in marketing models as well as how to solve the problem of information deficiency. Among those models, Finite Mixture and hierarchical Bayes models are considered superior. However, both methods still have their weakness in estimating reliable individual parameters. In order to verify empirically that combining information about individuals with similar behavioral pattern will greatly increase predictive validity, we combine market segmentation and hierarchical Bayes methods to propose a new hierarchical Bayes conjoint segmentation model.