Online Platform Market Reputation Systems. Text Analytics and Machine Learning Approaches to Online Reputation

      Аннотация: This research investigates the relationship among text review score, review length, agent behavior, and transfer-learning applications by constructing a quantitative measure with Airbnb online text reviews. The first essay focuses on studying the effects of review length on text review score. It constructs numerical text review scores by applying text analytics and machine learning techniques to more than three million user-generated text reviews on Airbnb. With the text review scores, this essay analyzes the impact of review length on text review scores and obtains insights on interplay relationship among text review score, review length, review age, and active reviewer. It finds an inverted U-shaped relation between review length and text review scores and a long-term downward trend of text review length cross all online platforms. This research contributes to online reputation field by building an innovative text review reputation measure and a large online labelled review corpus (i.e., Airbnb review knowledge base), investigating effects of review length on text review scores, and revealing a long-term trend of the online platform review length.

Ключевые слова: Internet platform, sharing economy, online trust and reputation systems, text review score, review length, text analytics, AI, Machine Learning, online purchase decision making.