Quality assessment of answers with user-identified criteria and data-driven features in social Q&amp;A

Fu, Hengyi; Oh, Sanghee

doi:10.1016/j.ipm.2018.08.007

Cited by 60 publications

(43 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Along with this, Bartlett's Sphericity test result (p<0.05) reveals that there is a significant connection between variables in the scale. This result also demonstrates that the data are suitable for factor analysis (Hutcheson & Sofroniou, 1999as cited in Field, Miles, & Field, 2012 In the first step of the factor analysis, which was performed, it was seen that the eigenvalues of 15 factors were above 1 (Fu & Oh, 2019;Yu et al, 2015) (See Figure 1). However, when the factor and loading rates are examined, it is observed that for many factors, there were very few of the items with a load value above 0.40, or in some cases no-load value for some factors.…”

Section: Exploratory Factor Analysis and Item-total Correlations Resultsmentioning

confidence: 56%

“…During the exploratory factor analysis, items were excluded from the scale when they had factor load values below 0.40 (Erfanmanesh et al, 2012;Fu & Oh, 2019;O'Brien & Toms, 2010 and/or the items with a high load value of more than one factor, and when the difference between the load values of these factors was less than 0.1. As a result of this, the 6 th , 16 th , 18 th , 47 th , 52 nd , 63 rd , 65 th , and 73 rd items were excluded from the scale because their factor load factor was below 0.40; the 8 th , 10 th , 12 th , 14 th , 21 st , 24 th , 25 th , 26 th , 27 th , 28 th , 29 th , 33 rd , 35 th , 36 th , 37 th , 39 th , 40 th , and 43 rd items were excluded from the scale because they had more than one load values where the difference between the load values was less than 0.10.…”

Section: Exploratory Factor Analysis and Item-total Correlations Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Bilgi Arama Süreci Tutum Ölçeği (BASTÖ): Güvenirlik ve Geçerlik Çalışması

Çakmak¹,

Baysen²

2019

Turk Kutuphaneciligi - Turkish Librarianship

View full text Add to dashboard Cite

The main aim of this study is to develop a valid and reliable measurement tool, which can reveal the attitudes of undergraduate students regarding the Information Search Process (ISP). Internal consistency and reliability computed with Cronbach Alpha (n=374) and test-retest. Construct validity was calculated through factor analyses, Exploratory Factor Analysis -EFA (n=883), and Confirmatory Factor Analysis-CFA (n=200). Additionally, varimax rotation, item-total correlations, and interdimensional correlations were conducted. After conducting these tests reliable and valid attitude scale, Attitude Scale for the Information Search Process (ASISP) with a total of 46 items, and a five-dimensional structure measuring individuals' attitudes toward ISP was finalized. Thus, ASISP is constituted by, "task initiation," "topic selection process," "defining focused topic", "information collection and search closure," and the "skills of research methodology" dimensions. The results of the study support the idea that the ISP is culture depended.

show abstract

Section: Exploratory Factor Analysis and Item-total Correlations Resultsmentioning

confidence: 56%

Section: Exploratory Factor Analysis and Item-total Correlations Resultsmentioning

confidence: 99%

Bilgi Arama Süreci Tutum Ölçeği (BASTÖ): Güvenirlik ve Geçerlik Çalışması

Çakmak¹,

Baysen²

2019

Turk Kutuphaneciligi - Turkish Librarianship

View full text Add to dashboard Cite

show abstract

“…For example, Shah and Pomerantz (Shah & Pomerantz, 2010) asked workers to rate the quality of each answer based on 13 criteria used to derive the quality criteria. Fu and Oh (2019) identified user criteria and data‐driven features for assessing the quality of the website answers. Liu, Feng, Liu, Hu, and Wang (2015) used co‐training to predict the quality of user‐generated answers.…”

Section: Introductionmentioning

confidence: 99%

“…These two types of evaluations are different. In the evaluation of answers in specific Q&A documents, the length, relevance and ratings are important factors (Fu & Oh, 2019). However, during the overall evaluation of cQA websites, we focused more on macro aspects, such as relevance to popular topics and the coverage of a wide range of fields.…”

Section: Introductionmentioning

confidence: 99%

A hybrid MCDM model combining DANP with TODIM to evaluate the information quality of community question answering in a two‐dimensional linguistic environment

Peng

et al. 2020

Expert Systems

View full text Add to dashboard Cite

Evaluating the information quality of cQA (community question‐answering) websites helps users select a cQA website with high‐quality information and improves the information quality. In this paper, an approach to evaluating the information quality of cQA based on a novel hybrid multicriteria decision‐making (MCDM) model is proposed. First, the source, content, expression and usefulness criteria for evaluating the information quality of the cQA are established. Then, considering the ambiguity and inner correlations of the criteria, DANP ([DEMATEL]‐based Analytic Network Process) and TODIM (interactive and multiple attribute decision making, in Portuguese) methods are combined in a two‐dimensional linguistic environment to process linguistic evaluation information. Lastly, the proposed approach is applied to evaluate the quality of five popular cQA websites. The key criteria are identified, and the evaluation results are derived comprehensively. The key factors consist of reputation, coverage, politeness, usability, helpfulness, clarity, readability and conciseness. The websites, Know almost and Baidu knows, perform better in terms of information quality. The application, along with the sensitivity analysis and comparative analysis, shows the effectiveness of the proposed model. The proposed approach has both practical and research implications. It provides an approach for users to choose websites, and, for operators, to improve the information quality of their website. The evaluation criteria and their relations provide a reference for research on the information quality. The evaluation method considers the user's psychological information to provide a more accurate MCDM approach. The comprehensive aspects of the experiment can be used to verify the other MCDM methods.

show abstract

“…Supervised machine learning is a typical one in dealing with certain problems namely, deep learning and statistical learning. End-to-end (e2e) Deep Neural Networks (DNN) could evade multifaceted characteristics of engineering and it includes improved learning ability when compared with previous learning approaches, therefore it has turn out to be a major flow in mining QA matching [8] [9]. Consequently, owing to the growing recognition of social Q&A [10] [11], a mixture of diverse questions is often required on Social Network Sites (SNSs).…”

Section: Introductionmentioning

confidence: 99%

Hybrid CNN-LSTM Model for Answer Identification

Moholkar¹,

Patil²

2019

IJRTE

View full text Add to dashboard Cite

User quest for information has led to development of Question Answer (QA) system to provide relevant answers to user questions. The QA task are different than normal NLP tasks as they heavily depend to semantics and context of given data. Retrieving and predicting answers to verity of questions require understanding of question, relevance with context and identifying and retrieving of suitable answers. Deep learning helps to produce impressive performance as it employs deep neural network with automatic feature extraction methods. The paper proposes a hybrid model to identify suitable answer for posed question. The proposes power exploits the power of CNN for extracting features and ability of LSTM for considering long term dependencies and semantic of context and question. Paper provides a comparative analysis on deep learning methods useful for predicting answer with the proposed method .The model is implemented on twenty tasks of babI dataset of Facebook .

show abstract

Quality assessment of answers with user-identified criteria and data-driven features in social Q&A

Cited by 60 publications

References 22 publications

Bilgi Arama Süreci Tutum Ölçeği (BASTÖ): Güvenirlik ve Geçerlik Çalışması

Bilgi Arama Süreci Tutum Ölçeği (BASTÖ): Güvenirlik ve Geçerlik Çalışması

A hybrid MCDM model combining DANP with TODIM to evaluate the information quality of community question answering in a two‐dimensional linguistic environment

Hybrid CNN-LSTM Model for Answer Identification

Contact Info

Product

Resources

About