Community Question Answering Sites Research Articles

Collusive spam detection from Chinese community question answering sites: A collective classification framework

With Community Question Answering (CQA) sites evolving into quite popular knowledge-sharing platforms on the Internet, they have also become ideal places for various spammers to spread fake or promotional information. Recently, with the rapid development of crowdsourcing systems, numerous malicious users have launched organized spam campaigns, conducting many spam accounts to carry out collusive spamming activities on CQA sites. In these campaigns, the spammers do not act independently but post deceptive questions and answers (Q&As) collaboratively, which makes the Q&As closely related to each other, but the spam clues of them are even less visible. Therefore, most existing spam detection works may fail to detect these carefully organized and posted collusive CQA spam. In this paper, taking Baidu Zhidao, a popular CQA platform in Chinese, as the study object, we propose a Collective Classification framework for community Question Answering spam detection (CCQA), which collectively identifies the collusive CQA spam using Q&A features and the correlations among Q&As. First, we define the Deceptive Pattern of Q&As, based on which the real Q&A groups are extracted. Then, we extract several highly discriminative Q&A features from both individual and group levels, and propose several types of correlations, which correlate the Q&As that are more likely to have the same labels. After uniformly modeling the Q&As, features, and correlations in the Attributed Heterogeneous Information Network (AHIN), a semi-supervised collective classification algorithm is proposed to detect the collusive Q&A spam. Experimental results on a real-life dataset demonstrate that CCQA can accurately detect the collusive CQA spam, and outperform a number of competitive baselines.

Finding Experts in Community Question Answering System Using Trie String Matching Algorithm with Domain Knowledge

In recent days, different community question-answering systems have been evolving as online forums for knowledge sharing among users across the world. However, finding experts on the questions is a key issue in the community question-answering system. To address this issue, a novel approach is proposed in this paper for finding experts from the community question-answering website using an exact string-matching algorithm along with domain knowledge. The proposed system is designed with three phases: i) User Profile Modeling, ii) Question Preprocessing, and iii) Expert Recommendation. Initially, the matrix factorization model is adopted for user profile modeling where the tags and the answerer’s information are represented in the form of a matrix. Then, the domain-wise grouping of tags is done to minimize the search time of the tags. The question preprocessing is done using a keyword extraction algorithm to extract the keywords. Finally, the expert recommendation is accomplished using the trie string matching algorithm and key-value mapping process. For doing experiments, a stack overflow community question-answering website is utilized in this work. The performance of the system is measured and the results section proved that the system achieved 91.2% accuracy.

Question Tags or Text for Topic Modeling: Which is better

Topic modelling is a probabilistic based statistical model used to find the latent topics that best depicts the content of the documents. Community Question Answering websites such as Quora, Stack Overflow and Yahoo! Answers have been prevalently in use, performs topic modeling as lot of queries pour in on daily basis which make it challenging to understand, summarize and synthesize the main topic of discussions. On these websites there are basically two sources of information that are available to analyze the key latent topics: questions text and tags. Questions are in textual format and tags are the keywords or tokens that are related to the question being asked which describes the content of the question. In past studies, most of the researchers have used question text for the purpose of topic modeling. It is still unclear why tag is not being considered for topic modeling. To combat this issue, this paper performs topic modeling using both question tags and text. The topic modeling based on tags has been compared with text based on two metrics namely coherence and perplexity. Experiment has been conducted on three real time datasets namely Artificial intelligence, Software Engineering and quantum computing from Stack exchange website. At high level tag-based topic modelling looked promising but closer observation revealed the opposite. It has been found that topic modeling using question text is preferable as topic modelling using tags collapses after a certain number of topics.

Multi-perspective respondent representations for answer ranking in community question answering

Answer ranking is an important task in community question answering (CQA) systems. It aims at ranking useful answers above useless answers. Existing works learn respondents’ expertise to help estimate qualities of answers. However, in most of these works, the expertise is only learned from the history answers. As a result, structure correlations between question raisers and respondents are usually ignored. Besides, these works lack an efficient way to learn respondent expertise from extensive history answers. To address the limitations, we propose a novel multi-perspective respondent representation learning(MPRR) network. First, our model learns embeddings of raisers and respondents through a heterogeneous information network(HIN) constructed by the answering records in CQA websites. The structure correlations between raisers and respondents are preserved in the learned embeddings. Second, a freezed pre-trained language model is used to learn respondents’ expertise from history answer contents more quickly. Then the multi-perspective respondent representations are generated based on their expertise and the embeddings learned in the HIN. At last, the raisers, respondents, questions, and answers are all considered to compute the matching scores. We evaluate our model on three real-world CQA dataset. Experiment results show that MPRR outperforms all baseline models with three ranking metrics on all datasets.

What Do Users Think of Promotional Gamification Schemes? A Qualitative Case Study in a Question Answering Website

In recent years, studies on the user experience have emerged as an indispensable part of any gamification research. The study of user experience enables gamification designers and practitioners to design or adapt their gamification schemes in a more knowledgeable and efficacious manner. However, one popular gamification scheme that has largely remained under-researched in terms of user experience is promotional gamification, which refers to an optional and time-limited gamification program that usually mounts an already gamified platform to increase user incentive and engagement for a short span of time (e.g., during the holiday season). The current study undertakes the first steps necessary to explore users' experiences of working with a promotional gamification scheme in a large-scale online community. To this end, we conduct an extensive qualitative case study of users' experiences with a promotional gamification scheme on the Community Question Answering Website (CQA) of Stack Exchange, called Winter Bash (WB). Notably, the purpose of WB is to operate as a makeshift solution that prevents the decline in user contributions during the holiday season. However, like many other gamification schemes, WB is not devoid of issues, and our research helps identify those issues without overlooking the WB's strengths. Our study denotes not only the first (empirical) typology of users' affective responses to promotional gamification schemes but also the first classification of (de)motivational factors involved in user engagement. At its core, this study comprises two salient parts: (1) a content analysis of user-generated data regarding WB (from the past eight years), and (2) a series of semi-structured interviews with 17 international users who are familiar with WB. We triangulate our findings from (1) and (2) by performing a similar content analysis for two other promotional gamification schemes, namely "Answerathon" (from Travel Meta) and "Discussion Tournament" (from Reddit). Based on the findings of this study, we present certain guidelines for gamification designers and practitioners, enabling them to deploy or adapt their promotional gamification schemes in a more knowledgeable and effective manner. Finally, our work is concluded by highlighting a few novel research opportunities for researchers invested in the fields of Human-Computer Interaction (HCI) and Computer-Supported Cooperative Work (CSCW).

EXPERT RECOMMENDATION THROUGH TAG RELATIONSHIP IN COMMUNITY QUESTION ANSWERING

Community Question Answering (CQA) services are technical discussion forums websites on social media that serve as a platform for users to interact mainly via question and answer. However, users of this platform have posed dissatisfaction over the slow response and the preference for user domains due to the overwhelming information in CQA websites. Numerous past studies focusing on expert recommendation are solely based on the information available from websites where they rarely account for the preference of users’ domain knowledge. This condition prompts the need to identify experts for the questions posted on community-based websites. Thus, this study attempts to identify ranking experts’ derived from the tag relationship among users in the CQA websites to construct user profiles where their interests are realized in the form of tags. Experts are considered users who post high-quality answers and are often recommended by the system based on their previous posts and associated tags. These associations further describe tags that often co-occur in posts and the significant domains of user interest. The current study further explores this relationship by adopting the “Tag Relationship Expert Recommendation (TRER)” method where Questions Answer (QA) Space is utilized as a dataset to identify users with similar interests and subsequently rank experts based on the tag-tag relationship for user’s question. The results show that the TRER method outperforms existing baseline methods by effectively improving the performance of relevant domain experts in CQA, thereby facilitating the expert recommendation process in answering questions posted by technical and academic professionals.

Feature-based question routing in community question answering platforms

Community question answering (CQA) platforms are receiving increased attention and are becoming an indispensable source of information in different domains ranging from board games to physics. The success of these platforms dependent on how efficiently new questions are assigned to community experts, known ascalled question routing. In this paper, we address the problem of question routing by adopting a learning to rank approach over five CQA websites in the context of which we introduce 74 features and systematically classify them into content-based and social-based categories. Our extensive experiments on datasets from five real online question answering websites indicate that content-based features related to tags and topics as well as social features that are related to user characteristics and user temporality are effective for question routing. Our work shows the ability to improve performance compared to the state-of-the-art neural matchmaking methods that lack the interpretability offered by our work. The improvement can be as high as on average 2.47% and 1.10% in terms of common ranking metrics, Normalized Discounted Cumulative Gain (NDCG) and Mean Average Precision (MAP) respectively, compared to our best baselines.

Knowledge-enhanced attentive learning for answer selection in community question answering systems

In a community question-answering (CQA) system, the answer selection task is used to identify the best answer for a specific question. This plays a key role in improving service quality by recommending appropriate answers to new questions. Recent advances in CQA answer selection have focused on enhancing performance by incorporating community information, and particularly the expertise (previous answers) and authority (position in the social network) of a respondent. However, existing approaches to incorporating this information are limited, as they (a) consider either the expertise or the authority, but not both; (b) ignore domain knowledge that could differentiate between the topics of previous answers; or (c) simply use authority information to adjust the similarity score, rather than fully integrating it into the process of measuring the similarity between the question and answer segments. We propose a new approach called the knowledge-enhanced attentive answer selection (KAAS) model, which enhances performance by (a) considering both the expertise and the authority of the answerer; (b) utilizing human-labeled tags, a taxonomy of tags, and votes as domain knowledge to infer the expertise of the respondent; (c) using a matrix decomposition of the social network (based on ‘following’ relationships) to infer the authority of the respondent and incorporating this information into the process of evaluating the similarity between segments. In addition, we incorporate an external knowledge graph to capture more professional information for CQA systems for vertical communities. We also adopt an attention mechanism to integrate our analysis of both questions and answers texts and the aforementioned community information. Experiments with both vertical and general CQA sites demonstrate the superior performance of the proposed KAAS model.

A Collaborative AI-Enabled Pretrained Language Model for AIoT Domain Question Answering

Large-scale knowledge in the Artificial Intelligence of Things (AIoT) field urgently needs effective models to understand human language and automatically answer questions. Pre-trained language models (PLMs) achieve state-of-the-art performance on some question answering (QA) datasets, but few models can answer questions on AIoT domain knowledge. Currently, AIoT domain lacks sufficient QA datasets and large-scale pre-training corpora. We propose RoBERTa-AIoT to address the problem of the lack of high-quality large-scale labeled AIoT QA datasets. We construct an AIoT corpus to further pre-train RoBERTa and BERT. RoBERTa-AIoT and BERT-AIoT leverage unsupervised pre-training on a large corpus composed of AIoT-oriented Wikipedia webpages to learn more domain-specific context and improve performance on the AIoT QA tasks. To fine-tune and evaluate the model, we construct 3 AIoT QA datasets based on the community QA websites. We evaluate our approach on these datasets and the experimental results demonstrate the significant improvements of our approach.

A Data-Driven Approach to Question Subjectivity Identification in Community Question Answering

Automatic Subjective Question Answering (ASQA), which aims at answering users'subjective questions using summaries of multiple opinions, becomes increasingly important. One challenge of ASQA is that expected answers for subjective questions may not readily exist in the Web. The rising and popularity of Community Question Answering (CQA) sites, which provide platforms for people to post and answer questions, provides an alternative to ASQA. One important task of ASQA is question subjectivity identification, which identifies whether a user is asking a subjective question. Unfortunately, there has been little labeled training data available for this task. In this paper, we propose an approach to collect training data automatically by utilizing social signals in CQA sites without involving any manual labeling. Experimental results show that our data-driven approach achieves 9.37% relative improvement over the supervised approach using manually labeled data, and achieves 5.15% relative gain over a state-of-the-art semi-supervised approach. In addition, we propose several heuristic features for question subjectivity identification. By adding these features, we achieve 11.23% relative improvement over word n-gram feature under the same experimental setting.

Intern retrieval from community question answering websites: A new variation of expert finding problem

In recent years, with the advent of new scientific and technical areas, hiring suitable interns has become very important for large companies. In internship programs, to reduce the risk of unsuccessful recruitment, companies are actively seeking those candidates who are ready to become skillful experts, and also incur the least financial costs. Therefore, generalists (i.e. Hyphen-shaped people) are the most suitable candidates for such positions. These candidates have general knowledge in the required skills of the position and do not have expertise in any other area. One of the environments that accurately reflect the knowledge of people is community question answering (CQA). This study is the first that focuses on retrieving interns from CQA websites. It uses the concepts of generalist and shape of expertise to identify suitable candidates for the internship programs of companies. Specifically, in this paper, we define the intern retrieval problem, in which given a set of required skills of an internship position, a ranking of candidates is generated so that the generalists who have general knowledge in those skills are retrieved in top ranks. We propose two retrieval models to address the problem. In order to evaluate the performance of these models, we introduce two specific measures (i.e. coverage and optimality). Our experiments on three test collections extracted from StackOverflow demonstrate the effectiveness of our models in comparison with several baselines.

Deep Neural Network to Predict Answer Votes on Community Question Answering Sites

Stack Exchange (SE) is a popular community question answering site where a large number of questions and answers are posted every day. Three tabs called active, oldest, and votes are used by SE to list the users’ answers. The default tab to display the answers is the votes’ tab. The problem with the current listing mechanism of the answer is that newly posted answers are always placed at the bottom of the list because of no votes. To receive the votes from the users, the answer needs to be seen by the user. However, the current mechanism favoring the oldest answers to gain the user’s votes. To resolve this bias, and provide an equal opportunity to all answers, this paper suggested a deep learning-based framework that predicts virtual votes for answers as soon as it is being posted on CQAs. The predicted votes may use to lists the answers on SE. The proposed model helps the users to know how many votes may be received their answer in the future. This also motivates the users to post high-quality answers to receive more votes. The prediction of votes required only the textual content of the answers and hence it is free from the handcrafted feature engineering. To validate the model, three different datasets of the SE are used and found a promising performance on each dataset.

Evaluating community question-answering websites using interval-valued intuitionistic fuzzy DANP and TODIM methods

As important knowledge-sharing platforms, community question-answering (CQA) websites have attracted the attention of many users. An evaluation of the quality of CQA websites can assist users in selecting high-quality CQA websites and can help the operators of CQA websites achieve the most beneficial improvements. An approach to the evaluation of CQA websites is proposed in this study. First, a CQA service quality model is constructed to evaluate the quality of CQA websites based on the E-SERVQUAL model. A novel multicriteria decision-making (MCDM) model is proposed to address the ratings. In the model, uncertainty and consistency are combined in an interval-valued intuitionistic fuzzy (IVIF) environment to determine two kinds of expert weights. The DANP method, along with experts’ weights, is extended to the IVIF environment to derive the criteria weights. The IVIF TODIM method, along with experts’ weights, is used to rank the CQA websites in terms of quality. Finally, the proposed approach is applied to evaluate the quality of five CQA websites. The experimental results of the application, along with sensitivity analysis and comparative analysis, demonstrate the feasibility and practicality of the proposed approach.

A Novel Expert Finding System for Community Question Answering

With the popularity of community question answering (CQA) sites, the research on identifying the expert users in online communities attracted increasing attention. We present a novel expert ranking algorithm based on the quality of user posts and the authority of user in community, and the similarity between the knowledge tags of users and questions in CQA sites is adopted in our scheme. Experimental results show that our scheme has better performance and accuracy under the same background with an amount of data samples.

KNOWLEDGE-ENHANCED ATTENTIVE LEARNING FOR ANSWER SELECTION IN COMMUNITY QUESTION ANSWERING SYSTEMS

In the community question answering (CQA) system, the answer selection task aims to identify the best answer for a specific question, and thus is playing a key role in enhancing the service quality through recommending appropriate answers for new questions. Recent advances in CQA answer selection focus on enhancing the performance by incorporating the community information, particularly the expertise (previous answers) and authority (position in the social network) of an answerer. However, existing approaches for incorporating such information are limited in (a) only considering either the expertise or the authority, but not both; (b) ignoring the domain knowledge to differentiate topics of previous answers; and (c) simply using the authority information to adjust the similarity score, instead of fully utilizing it in the process of measuring the similarity between segments of the question and the answer. We propose the Knowledge-enhanced Attentive Answer Selection (KAAS) model, which enhances the performance through (a) considering both the expertise and the authority of the answerer; (b) utilizing the human-labeled tags, the taxonomy of the tags, and the votes as the domain knowledge to infer the expertise of the answer; (c) using matrix decomposition of the social network (formed by following-relationship) to infer the authority of the answerer and incorporating such information in the process of evaluating the similarity between segments. Besides, for vertical community, we incorporate an external knowledge graph to capture more professional information for vertical CQA systems. Then we adopt the attention mechanism to integrate the analysis of the text of questions and answers and the aforementioned community information. Experiments with both vertical and general CQA sites demonstrate the superior performance of the proposed KAAS model.

Explanatory Q&A recommendation algorithm in community question answering

PurposeIn community question answering (CQA), people who answer questions assume readers have mastered the content in the answers. Nevertheless, some readers cannot understand all content. Thus, there is a need for further explanation of the concepts that appear in the answers. Moreover, the large number of question and answer (Q&A) documents make manual retrieval difficult. This paper aims to alleviate these issues for CQA websites.Design/methodology/approachIn the paper, an algorithm for recommending explanatory Q&A documents is proposed. Q&A documents are modeled with the biterm topic model (BTM) (Yan et al., 2013). Then, the growing neural gas (GNG) algorithm (Fritzke, 1995) is used to cluster Q&A documents. To train multiple classifiers, three features are extracted from the Q&A categories. Thereafter, an ensemble classification model is constructed to identify the explanatory relationships. Finally, the explanatory Q&A documents are recommended.FindingsThe GNG algorithm shows good clustering performance. The ensemble classification model performs better than other classifiers. The both effect and quality scores of explanatory Q&A recommendations are high. These scores indicate the practicality and good performance of the proposed recommendation algorithm.Research limitations/implicationsThe proposed algorithm alleviates information overload in CQA from the new perspective of recommending explanatory knowledge. It provides new insight into research on recommendations in CQA. Moreover, in practice, CQA websites can use it to help retrieve Q&A documents and facilitate understanding of their contents. However, the algorithm is for the general recommendation of Q&A documents which does not consider individual personalized characteristics. In future work, personalized recommendations will be evaluated.Originality/valueA novel explanatory Q&A recommendation algorithm is proposed for CQA to alleviate the burden of manual retrieval and Q&A overload. The novel GNG clustering algorithm and ensemble classification model provide a more accurate way to identify explanatory Q&A documents. The method of ranking the explanatory Q&A documents improves the effectiveness and quality of the recommendation. The proposed algorithm improves the accuracy and efficiency of retrieving explanatory Q&A documents. It assists users in grasping answers easily.

Using cognition to resolve duplicacy issues in socially connected healthcare for smart cities

Social web has transformed healthcare communication as patients reach out to seek support and advice by connecting with other patients, caregivers and healthcare professionals. The influx of health-related queries and the volume of answers within the medical forums is a testimony to this adaption. The scalability, natural interaction and dynamism of the continuously collected and connected user-generated social big data can support health assessment, intervention and provisioning to produce the best kind of cognitive smart city. On the flip side the use of social media for healthcare communication suffers from data deluge, lack of reliability and quality, confidentiality and privacy (location/personal) issues. Duplicate questions, that is, queries with similar semantics (meaning) corrupt the filtering mechanism, increase the response time and compromise with the quality of the answer too. This research puts forward solutions to resolve the key challenge of duplicacy within the medical community Question-Answering sites (Medical CQAs). We propose to solve the semantic question matching problem for duplicate question pair detection, using a hybrid deep learning model, which combines a Co-attention based Bi-Directional Long Short-Term Memory (Bi-LSTM) Siamese neural network and a Multi-layer perceptron classifier to output the probability of a similarity match between the two questions. Euclidean distance function is then used to compute the similarity between questions. The proposed model is validated on 100 question pairs which are scrapped from three featured groups, namely, ‘Irritable Bowel Syndrome’, ‘Anxiety Disorder’ and ‘Menopause’ of Patient.info community forum and an accuracy of 86.375% is observed. The results obtained are comparable to that of the Quora’s state-of-the-art results for duplicate detection.

Identifying Experts in Community Question Answering Website Based on Graph Convolutional Neural Network

High-quality answers, usually given by experts, play an important role in community question answering (CQA) websites. Therefore, experts in these websites are defined as those who provide high-quality answers. We propose two semi-supervised learning models based on the graph convolution neural network (GCN) to identify them. Both models comprehensively extract features from the social behavior network, user profiles, and question and answer text. Specifically, we construct a social behavior network according to the co-answering relationship among answerers, which means every two answerers are connected if they answer the same questions. The difference between these two models is the methods of extracting text features. One model named GCN-Doc uses Doc2vec to get text vectors before training. The other model named GCN-Lstm with a long short term memory (LSTM) network extracts text features while training. Experiments using real-world data from Zhihu.com, one of the largest Chinese CQA websites, show that both GCN-Doc and GCN-Lstm can identify experts effectively comparing with baselines of PageRank and other GCN based neural network models. Besides, GCN-Lstm performs better than GCN-Doc.

Dynamic user modeling for expert recommendation in community question answering

Community Question Answering (CQA) websites has become an important channel for people to acquire knowledge. In CQA, one key issue is to recommend users with high expertise and willingness to answer the given questions, i.e., expert recommendation. However, a lot of existing methods consider the expert recommendation problem in a static context, ignoring that the real-world CQA websites are dynamic, with users’ interest and expertise changing over time. Although some methods that utilize time information have been proposed, their performance improvement can be limited due to fact that they fail they fail to consider the dynamic change of both user interests and expertise. To solve these problems, we propose a deep learning based framework for expert recommendation to exploit user interest and expertise in a dynamic environment. For user interest, we leverage Long Short-Term Memory (LSTM) to model user’s short-term interest so as to capture the dynamic change of users’ interests. For user expertise, we design user expertise network, which leverages feedback on users’ historical behavior to estimate their expertise on new question. We propose two methods in user expertise network according to whether the dynamic property of expertise is considered. The experimental results on a large-scale dataset from a real-world CQA site demonstrate the superior performance of our method.

Extracting core questions in community question answering based on particle swarm optimization

PurposeA large number of questions are posted on community question answering (CQA) websites every day. Providing a set of core questions will ease the question overload problem. These core questions should cover the main content of the original question set. There should be low redundancy within the core questions and a consistent distribution with the original question set. The paper aims to discuss these issues.Design/methodology/approachIn the paper, a method named QueExt method for extracting core questions is proposed. First, questions are modeled using a biterm topic model. Then, these questions are clustered based on particle swarm optimization (PSO). With the clustering results, the number of core questions to be extracted from each cluster can be determined. Afterwards, the multi-objective PSO algorithm is proposed to extract the core questions. Both PSO algorithms are integrated with operators in genetic algorithms to avoid the local optimum.FindingsExtensive experiments on real data collected from the famous CQA website Zhihu have been conducted and the experimental results demonstrate the superior performance over other benchmark methods.Research limitations/implicationsThe proposed method provides new insight into and enriches research on information overload in CQA. It performs better than other methods in extracting core short text documents, and thus provides a better way to extract core data. The PSO is a novel method used for selecting core questions. The research on the application of the PSO model is expanded. The study also contributes to research on PSO-based clustering. With the integration of K-means++, the key parameter number of clusters is optimized.Originality/valueThe novel core question extraction method in CQA is proposed, which provides a novel and efficient way to alleviate the question overload. The PSO model is extended and novelty used in selecting core questions. The PSO model is integrated with K-means++ method to optimize the number of clusters, which is just the key parameter in text clustering based on PSO. It provides a new way to cluster texts.