東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Modeling Text-based Search Behavior:...

Liu, Jia.

FindBook

Google Book

Amazon

博客來

Modeling Text-based Search Behavior: Linking User Online Queries with Content Preferences.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Modeling Text-based Search Behavior: Linking User Online Queries with Content Preferences./
作者:	Liu, Jia.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2017,
面頁冊數:	141 p.
附註:	Source: Dissertation Abstracts International, Volume: 78-09(E), Section: A.
Contained By:	Dissertation Abstracts International78-09A(E).
標題:	Marketing. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10267726
ISBN:	9781369707595

Modeling Text-based Search Behavior: Linking User Online Queries with Content Preferences.
Liu, Jia.

Modeling Text-based Search Behavior: Linking User Online Queries with Content Preferences. - Ann Arbor : ProQuest Dissertations & Theses, 2017 - 141 p.

Source: Dissertation Abstracts International, Volume: 78-09(E), Section: A.

Thesis (Ph.D.)--Columbia University, 2017.

This thesis consists of three related essays which explore how users form online search queries, and how to estimate users' content preferences and their dynamics from web search behaviors. In Essay I, we explore the link between users' content preferences and their search queries. The existence of semantic relationships between queries and results differentiates query formation from traditional, discrete-choice based search. Accordingly, our research questions are as follows: (i) Under which conditions is it more beneficial for users to leverage the semantic relationships between queries and results when formulating queries? (ii) Are users able to leverage semantic relationships when formulating queries? (iii) How should researchers represent these semantic relationships? (iv) What are users' beliefs on these semantic relationships? We find that leveraging semantic relationships is particularly useful to users in retrieving content that is aligned with their preferences, when these preferences are less specific and when it is costly to formulate longer queries. We also find that users have the ability to formulate queries that leverage semantic relationships. The set of semantic relationships, which capture the probability that any query will activate any set of words, grows exponentially with the vocabulary size. Fortunately, we show they may be approximated by a set of activation probabilities at the word level, which grows only polynomially. We find that users' beliefs on these relationships are biased upwards, and that they are not asymmetric enough.

ISBN: 9781369707595Subjects--Topical Terms:

536353
Marketing.

Modeling Text-based Search Behavior: Linking User Online Queries with Content Preferences.
LDR:04682nmm a2200313 4500 001 2154586
005 20180419104822.5
008 190424s2017 ||||||||||||||||| ||eng d
020 $a 9781369707595
035 $a (MiAaPQ)AAI10267726
035 $a (MiAaPQ)columbia:13855
035 $a AAI10267726
040 $a MiAaPQ $c MiAaPQ
100 1 $a Liu, Jia. $3 2095876
245 1 0 $a Modeling Text-based Search Behavior: Linking User Online Queries with Content Preferences.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2017
300 $a 141 p.
500 $a Source: Dissertation Abstracts International, Volume: 78-09(E), Section: A.
500 $a Adviser: Olivier Toubia.
502 $a Thesis (Ph.D.)--Columbia University, 2017.
520 $a This thesis consists of three related essays which explore how users form online search queries, and how to estimate users' content preferences and their dynamics from web search behaviors. In Essay I, we explore the link between users' content preferences and their search queries. The existence of semantic relationships between queries and results differentiates query formation from traditional, discrete-choice based search. Accordingly, our research questions are as follows: (i) Under which conditions is it more beneficial for users to leverage the semantic relationships between queries and results when formulating queries? (ii) Are users able to leverage semantic relationships when formulating queries? (iii) How should researchers represent these semantic relationships? (iv) What are users' beliefs on these semantic relationships? We find that leveraging semantic relationships is particularly useful to users in retrieving content that is aligned with their preferences, when these preferences are less specific and when it is costly to formulate longer queries. We also find that users have the ability to formulate queries that leverage semantic relationships. The set of semantic relationships, which capture the probability that any query will activate any set of words, grows exponentially with the vocabulary size. Fortunately, we show they may be approximated by a set of activation probabilities at the word level, which grows only polynomially. We find that users' beliefs on these relationships are biased upwards, and that they are not asymmetric enough.
520 $a Based on the findings from Essay I, we then develop a practical method that can estimate user content preferences based on search queries in Essay II. Specifically, we develop an innovative topic model, Hierarchically Dual Latent Dirichlet Allocation (HDLDA), which not only identifies topics in search queries and webpages, but also how the topics in search queries relate to the topics in the corresponding top search results (which captures semantic-based search). Using the output from HDLDA, a user's content preferences may be estimated on the fly based on their search queries. We validate our proposed approach across different product categories using an experiment in which we observe participants' content preferences, the queries they formulate, and their browsing behavior. Our results suggest that our approach can help firms extract and understand the preferences revealed by users through their search queries, which in turn may be used to optimize the production and promotion of online content.
520 $a Consumers regularly interact with search engines to acquire information matching their changing content preferences, by entering search queries and clicking on the results. Hence, Essay III develops a contentbased dynamic search model (CDSM) that can collaboratively model users' text-based search behaviors (typing a query), subsequent discrete search behaviors (clicking on a link), and the content that users encounter during the search process, while also considering changes over time in users' preferences underlying these search behaviors. We also develop a variational inference algorithm for CDSM that allows for largescale model inference. Our proposed modeling framework provides several managerial applications. First, CDSM can interpret and explain query usage, clicks on URLs, and the relationship between the two. Second, our model can be used to recommend results for both search engines and marketers. We evaluate CDSM using Bing user search on popular TV shows. We study how user search behaviors and underlying content preferences vary across time (i.e., before, during, and after a show is aired).
590 $a School code: 0054.
650 4 $a Marketing. $3 536353
690 $a 0338
710 2 $a Columbia University. $b Business. $3 1681698
773 0 $t Dissertation Abstracts International $g 78-09A(E).
790 $a 0054
791 $a Ph.D.
792 $a 2017
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10267726