The Search Value Model: Detecting Abnormal Searching Behavior
Juan Luis Nicolau,Hyoeun Kim,Xianwei Liu
DOI: https://doi.org/10.1016/j.annals.2020.103007
IF: 12.853
2020-01-01
Annals of Tourism Research
Abstract:The search value model: detecting abnormal searching behavior Web search data reflect the attention people pay to particular topics (Li and Law, 2020).Searches for destinations, hotels, or airlines can be used as an explanatory variable for future firm sales (Bi et al.,2020).In fact, literature shows accuracy improvements when search engine data are included in tourism forecasting (Song et al., 2019; Li and Law, 2020).This article proposes a search value model to identify abnormal searching behavior, that is, significantly higher (or lower) search volume than expected on a topic for a specific period of time.To accomplish this objective, we first look at the market model.As a statistical model well established in financial economics, the market model is commonly used as a return-generating process (Sharpe, 1964).This model connects the return on a specific stock with that on a set of representative stocks, that is, the market portfolio (for which a common proxy are stock indexes, such as the S&P 500 Index (Gim and Jang, 2020)).The