Zhangxi Lin

Capturing the essence of word-of-mouth for social commerce: Assessing the quality of online e-commerce reviews by a semi-supervised approach

In e-commerce, online product reviews significantly influence the purchase decisions of buyers and the marketing strategies employed by vendors. However, the abundance of reviews and their uneven quality make distinguishing between useful and useless reviews difficult for potential customers, thereby diminishing the benefits of online review systems. To address this problem, we develop a semi-supervised system called Online Review Quality Mining (ORQM). Embedded with independent component analysis and semi-supervised ensemble learning, ORQM exploits two opportunities: the improvement of classification performance through the use of a few labeled instances and numerous unlabeled instances, and the effectiveness of the social characteristics of e-commerce communities as identifiers of influential reviewers who write high-quality reviews. Three complementary experiments on datasets from Amazon.com show that ORQM exhibits remarkably higher performance in classifying reviews of different quality levels than do other well-accepted state-of-the-art text mining methods. The high performance of ORQM is also consistent and stable even under limited availability of labeled instances, thereby outperforming other baseline methods. The experiments also reveal that (1) the social features of reviewers are important in deriving better classification results; (2) classification results are affected by product type given the different purchase habits of consumers; and (3) reviews are contingent on the inherent nature of products, such as whether they are search goods or experience goods, and digital products or physical products, through which purchase decisions are influenced.

Media-aware quantitative trading based on public Web information

Abstract Recent studies in behavioral finance discover that emotional impulses of stock investors affect stock prices. The challenge lies in how to quantify such sentiment to predict stock market movements. In this article, we propose a media-aware quantitative trading strategy utilizing sentiment information of Web media. This is achieved by capturing public mood from interactive behaviors of investors in social media and studying the impact of firm-specific news sentiment on stocks along with such public mood. Our experiments on the CSI 100 stocks during a three-month period show that a predictive performance in closeness to the actual future stock price is 0.612 in terms of root mean squared error, the same direction of price movement as the future price is 55.08%, and a simulation trading return is up to 166.11%.

A framework for exploring organizational structure in dynamic social networks

Recent research has provided promising results relating to discovering communities within a social network. We find that further representing the organizational structure of a social network is an interesting issue that helps gain better understandings of the social network. In this paper, we define a data structure, named Community Tree, to depict the organizational structure and provide a framework for exploring the organizational structure in a social network. In this framework, an algorithm, which combines a modified PageRank and Random Walk on graph, is developed to derive the community tree from the social network. In the real world, a social network is constantly evolving. In order to explore the organizational structure in a dynamic social network, we develop a tree learning algorithm, which employs tree edit distance as the scoring function, to derive an evolving community tree that enables a smooth transition between two community trees. We also propose an approach to threading communities in community trees to obtain an evolution graph of the organizational structure, by which we can reach new insights from the dynamic social network. The experiments conducted on synthetic and real dataset demonstrate the feasibility and applicability of the framework. Based on the theoretical outcomes, we further apply the proposed framework to explore the evolution of organizational structure with the 2001 Enron dataset, and obtain several interesting findings that match the context of Enron.

Effects of Borrower-Defined Conditions in the Online Peer-to-Peer Lending Market

In online Peer-to-Peer lending market, the borrower-defined conditions of loan requests predetermine the successfulness to receive loans. We analyze the transaction data of PPDai, a leading Peer-to-Peer lending market provider in China. By using the multinomial logit model to investigate the importance of borrowers’ decisions and their effects on funding results, we reveal that loan amount, acceptable maximum interest rate, and loan period decided by borrowers significantly influence the loan outcomes. For the unsuccessful listings, the requested loan amount has much more importance than other factors, while for the listings attracting more supply than the requested amount, the borrower’s acceptable maximum interest rate are more dominant than other factors to the outcomes. Besides, consistent to prior literature’s findings, PPDai borrower’s personal information and social capital also play major role in the transactions.

Creditworthiness Analysis in E-Financing Businesses - A Cross-Business Approach

To cope with the challenge of data scarcity in creditworthiness analysis for e-financing business, this paper proposes a cross-business analysis approach based on the assumption of behavior consistency for client in different e-commerce environments. By this approach we can analyze individuals’ creditworthiness by associating financial data on lending platforms and cross-business non-financial data on social media. We conceived three creditworthiness assessment models, and conduct the experimental study on Ant Financial Co-Creation Data Platform. The results verify that our cross-business creditworthiness analysis approach is effective.

Exploration of a Multi-dimensional Evaluation of Books Based on Online Reviews: A Text Mining Approach

With advancements made to the Internet, a considerable increase in the number and types of products available online has come. Yet, the large amount of online consumer reviews may present an obstacle to potential buyers. This study proposes a four-dimensional book evaluation system for use by leading online booksellers, thereby enabling potential buyers to form decisions based on differentiated criteria. This book evaluation system was empirically examined by employing a text mining approach and multivariate regression model. The findings here-in may aid in improving the understanding of the construction of online product evaluation systems.

An Idealet-Centric Scheme for Large Scale Open Innovation Systems

This paper intends to demonstrate how open innovation systems could be developed by tackling the challenging knowledge management problems that are encountered when aiming at involving very large audiences. This is the case when generalizing open innovation approach beyond companies to a wider societal context like in the case of national innovation systems. The Open Innovation Banking System (OIBS) project, funded by the European Social Fund (ESF) and the participating higher education institutions in Finland, is used as a basis for our discussion. It specifically aims at bringing the largely underutilized creativity of students and senior citizens to play. Among several technologies to develop OIBS, mashups as hybrid web applications can play an important role in such constantly evolving system and contents. However, relying only on unstructured text inputs, the services of textual content sharing for OIBS would require intelligent text processing that far exceeds the capability of such applications. In this paper, we propose an “idealet”-centric solution for representing the data submitted by users, enabling concise description, refinement and linking of ideas as input for innovation processes. An idealet is defined as the core knowledge about an innovative idea. The relationships among idealets and essays can be represented in a semantic network in terms of their relationships. This scheme allows the mashup applications for OIBS to more effectively retrieve, process, extract, and deliver the most important knowledge from an ocean of information contributed by participating information composer, reviewers, and users. The paper also discusses how the idealet-centric approach can be employed for a functional open innovation system.