題目： Calibrating the Helpfulness of Online Product Reviews: An Iterative Bayesian Probability Approach
主講人： 郭迅華（清華大學(xué)）
時(shí)間：2017年12月26日（周二）上午10：00
地點(diǎn)：主樓418
主講人介紹：
2000年獲得清華大學(xué)管理信息系統(tǒng)專業(yè)學(xué)士學(xué)位及計(jì)算機(jī)科學(xué)與技術(shù)專業(yè)學(xué)士學(xué)位，2005年獲得清華大學(xué)管理科學(xué)與工程專業(yè)碩士學(xué)位和博士學(xué)位。現(xiàn)任清華大學(xué)經(jīng)濟(jì)偉德國際1946bv官網(wǎng)副教授，主要研究領(lǐng)域?yàn)楣芾硇畔⑾到y(tǒng)、電子商務(wù)、社會網(wǎng)絡(luò)、商務(wù)智能。講授課程包括管理信息系統(tǒng)、信息技術(shù)與組織、計(jì)算機(jī)系統(tǒng)原理、計(jì)算機(jī)網(wǎng)絡(luò)。學(xué)術(shù)論文發(fā)表于MIS Quarterly、Journal of MIS、Communications of the ACM、DecisionSciences、INFORMS Journal on Computing、Information Systems Journal、Journal of Information Technology、Information Sciences、Information & Management、Decision Support Systems、Computers in Human Behavior、ACM Transactions on Knowledge Discovery from Data等信息系統(tǒng)領(lǐng)域重要國際期刊，以及《管理科學(xué)學(xué)報(bào)》、《管理世界》、《中國管理科學(xué)》、《系統(tǒng)工程理論與實(shí)踐》等重要國內(nèi)期刊，作為負(fù)責(zé)人或骨干參與了多項(xiàng)國家自然科學(xué)基金項(xiàng)目和企業(yè)項(xiàng)目。曾獲得清華大學(xué)學(xué)術(shù)新秀、優(yōu)秀博士畢業(yè)生榮譽(yù)稱號。曾于2008年在德國RWTH Aachen University
做訪問學(xué)者以及在MIT斯隆偉德國際1946bv官網(wǎng)擔(dān)任國際教職研究員。現(xiàn)任國際信息系統(tǒng)協(xié)會中國分會（CNAIS）常務(wù)理事兼副秘書長，《信息系統(tǒng)學(xué)報(bào)》主編助理，Electronic Commerce Research、Journal of Global Information Management等國際學(xué)術(shù)雜志編委會成員。
內(nèi)容介紹：
Voting mechanisms are widely adopted for evaluating the quality and reputation of user generated content, such as online product reviews. For the reviews that do not receive sufficient votes, techniques and models are developed to automatically assess their helpfulness levels. Existing methods are mostly centered on feature analysis, ignoring the information conveyed in the frequencies and patterns of user votes. Consequently, the accuracy of helpfulness measurement is limited. Inspired by related findings from prediction theories and consumer behavior research, we propose a novel approach characterized by the technique of iterative Bayesian distribution estimation, aiming to more accurately measure the helpfulness levels of reviews used for training prediction models. Using synthetic data and a real-world data set involving 1.67 million reviews and 5.18 million votes from Amazon, a simulation experiment and a two-stage data experiment show that the proposed approach outperforms existing methods on accuracy measures. Furthermore, an out-of-sample user study is conducted on Amazon Mechanical Turk as well as in a university lab. The results further illustrate the predictive power of the new approach. Practically, the research contributes to e-commerce by providing an enhanced method for exploiting the value of user-generated content. Academically, we contribute to the design science literature with the novel approach that may be adapted to a wide range of research topics such as recommender systems and social media analytics.
（承辦：管理工程系，科研與學(xué)術(shù)交流中心）
