虚假评论检测技术综述

随着互联网的发展,用户倾向于在购物、旅游、用餐之前参考线上评论.之后,他们也会发表评论来表达自身意见.线上评论越来越具有价值.评论对用户决策的重要导向作用催生了虚假评论.虚假评论,指用户由于利益、个人偏见等因素发布的不符合产品真实特性的评论.这些虚假评论语言上模仿真实评论,消费者很难识别出来.国内外学者综合运用自然语言处理技术来研究虚假评论检测问题.从特征工程的角度分析,虚假评论检测方法可以分为三类：基于语言特征和行为特征的方法、基于图结构的方法、基于表示学习的方法.主要描述了检测的一般流程,归纳了三类研究方法常用的特征,比较了方法的优缺点,并且介绍了研究常用的数据集.最后探讨了未来研究方向.     

基于文本和用户信息的在线评论质量检测

随着互联网的迅速发展,越来越多的用户评论出现在社交网站上。面对迅速增长的评论数据,如何为阅读评论的消费者提供准确、真实的高质量评论就显得尤为重要。评论质量检测旨在判断在线评论的质量,在传统的研究中,文本信息通常独立地被用于预测评论质量。但是在社交媒体上,每个文本之间不是独立的,而是可以通过发表文本的作者与其他文本相关联,即同一个用户或相近的用户发表的评论质量具有一定的相似性。因此,为了更好的构建文本的表示和研究文本之间基于用户的关联,该文基于神经网络模型分别构建用户和文本的表示,同时,为了放大用户信息的作用,我们进一步将基于注意力机制的用户信息融合到文本中,从而提高文本评论质量检测的效果。在Yelp 2013数据集上进行实验的结果表明,该模型能有效地提高在线评论质量检测的性能。     

基于主题模型和情感分析的垃圾评论识别方法研究

随着电子商务的飞速发展,网络购物越来越被消费者认同,而随之产生的产品评论给消费者的购买决策带来了影响。产品评论是指用户在购物站点上对商品的评价信息,而 经过分析和研究发现这些评论中充斥着大量的垃圾评论,因此垃圾评论的识别成了电子商务在提高服务质量的过程中需解决的重要问题之一。根据垃圾评论的主要特点提出LDA-SP(LDA-Sentiment Polarity)垃圾评论识别方法。首先利用LDA主题模型过滤出内容型垃圾评论,然后结合情感分析识别出欺骗型垃圾评论。对网络商城的大量评论数据进行准确度分析实验的结果表明,LDA-SP方法的识别准确度高于传统的LDA主题模型和单一的情感极性分析方法,能够有效地检测垃圾评论,从而使产品评论信息更加客观准确,为电子商务用户提供了有效的参考信息。     

一种Web评论自动抽取方法

Web用户评论是许多重要应用的信息来源,比如公众舆情的检测与分析,Web用户评论必须从网页中准确地抽取出来.用户生成内容(user-generated content)不受页面模板的限制,这就给Web数据抽取提出了新的挑战:首先,不同用户评论内容的不一致性严重影响了评论记录在DOM树和视觉上的相似性;其次,评论内容在DOM树中是一棵复杂的子树,而且彼此之间在DOM树中的结构相差巨大.为了解决这两个问题,提出了一种完整的解决方案,使用多种技术来实现对用户评论内容的抽取.抽取过程分为两个步骤,基于深度加权的树相似性算法评论记录首先从网页中抽取出来,然后通过比较DOM树中节点的一致性,将纯粹的用户评论内容从评论记录中抽取出来.在多个新闻网站和论坛网站上的实验结果表明,该方法可以达到较高的准确度和效率.     

面向用户需求获取的在线评论有用性分析

在线评论已成为互联网环境下用户需求获取的重要数据资源.然而,评论质量的良莠不齐严重干扰了需求挖掘的准确性和可信性.如何发现能够准确描述用户需求的有用评论是提高需求获取技术有效性的前提保障.针对这一问题,文中提出一种基于复杂网络的评论有用性分析方法,利用评论间的语义关联,从宏观的角度分析评论对于用户需求识别的有用程度,进而发现能够准确描述用户需求的评论.作者将评论看作一种内容互连的网络拓扑的形态,利用评论网络节点的重要性来度量评论的有用性,并通过拓扑势理论将用户的主观评价与网络拓扑结构的客观影响有机融合对评论网络节点重要性进行分析.实验结果表明,该方法所确定的高有用性评论能够保证用户需求获取具有较高的准确率和覆盖率.     

评论信息抽取技术在网络贸易系统平台上的应用

准确挖掘购物网站中的用户评论对商家和顾客进行有效的推荐具有重要作用.本文研究了现有的网络贸易系统平台的功能模块,给出了用户评论信息抽取的具体方法,并指出用户评论信息抽取模块应用到网络贸易系统平台的优势和实际意义.     

产品垃圾评论检测研究综述

互联网上的产品垃圾评论混淆视听,误导了潜在消费者。产品垃圾评论检测的目的就是将垃圾评论从评论文本中找到并去除,保留真实的产品评论供用户参考。首先将产品垃圾评论和互联网上其它常见的垃圾信息进行了对比,并把产品垃圾评论的检测和产品评论的质量判断、产品评论的情感分析等相关的工作进行了比较分析。然后从产品垃圾评论检测的数据集、检测方法两个角度对相关工作做了概述和分析。最后,在上述工作的基础上提出了一些产品垃圾评论检测研究中值得进一步关注的问题。     

基于用户和产品表示的情感分析和评论质量检测联合模型

情感分析旨在判断文本的情感倾向,而评论质量检测旨在判断评论的质量.情感分析和评论质量检测是情感分析中两个关键的任务,这两个任务受多种因素的影响而密切相关,同一个产品的情感倾向具有相似的情感极性;同时,同一个用户发表的评论质量也具有一定的相似性.因此,为了更好地研究情感分类和评论质量检测任务的相关性以及用户信息和产品信息分别对情感分类和评论质量检测的影响,提出了一个情感分析和评论质量检测联合模型.首先,使用深度学习方法学习评论的文本信息作为联系两个任务的基础;然后,将用户评论及产品评论作为用户的表示和产品的表示;在此基础上,采用用户注意力机制对用户的表示进行编码,采用产品注意力机制对产品的表示进行编码;最后,将用户表示和产品表示结合起来进行情感分析和评论质量检测.通过在Yelp2013和Yelp2015数据集上的实验结果表明,该模型与现有的神经网络模型相比,能够有效地提高情感分析和在线评论质量检测的性能.     

基于评论质量的多文档文本情感摘要

文本情感摘要任务旨在对带有情感的文本数据进行浓缩、提炼进而产生文本所表达的关于情感意见的摘要,用以帮助用户更好地阅读、理解情感文本的内容。该文主要研究多文档的文本情感摘要问题,重点针对网络上存在的同一个产品的多个评论进行摘要抽取。在情感文本中,情感相关性是一个重要的特点,该文将充分考虑情感信息对文本情感摘要的重要影响。同时,对于评论语料,质量高的评论或者说可信度高的评论可以帮助用户更好的了解评论中所评价的对象。因此,该文将充分考虑评论质量对文本情感摘要的影响。并且为了进行关于文本情感摘要的研究,该文收集并标注了一个基于产品评论的英文多文档文本情感摘要语料库。实验证明,情感信息和评论质量能够帮助多文档文本情感摘要,提高摘要效果。     

融合评论的多任务联合谣言检测方法

目前,针对微博领域的谣言检测方法主要基于微博正文,同时辅以用户评论特征、传播特征等信息进行判定。然而已有方法没有考虑用户评论质量会直接影响谣言检测的性能,质量低的评论甚至会引入无用甚至负面的特征,进而对谣言检测的性能带来更大的影响。针对该问题,基于用户评论和谣言检测的关联性,首次提出一种考虑评论有效性,并基于多任务联合学习的谣言检测方法。首先将谣言检测作为主任务,用户评论相关性检测为辅助任务;然后采用门控机制和注意力机制过滤和选择有效的用户评论特征;最后基于自主构建的3万条疫情微博谣言数据集进行实验。实验结果表明,对用户评论进行筛选不仅可以提升谣言检测性能,还能对用户评论质量进行判定。     

虚假评论识别研究综述

Web 2.0时代,消费者在在线购物、学习和娱乐时越来越多地依赖在线评论信息,而虚假的评论会误导消费者的决策,影响商家的真实信用,因此有效识别虚假评论具有重要意义。文中首先对虚假评论的范围进行了界定,并从虚假评论识别、形成动机、对消费者的影响以及治理策略4个方面归纳了虚假评论的研究内容,给出了虚假评论研究框架和一般识别方法的工作流程。然后从评论文本内容和评论者及其群组行为两个角度,对近十年来国内外的相关研究成果进行了综述,介绍了虚假评论效果评估的相关数据集和评价指标,统计分析了在公开数据集上实现的虚假评论有效识别方法,并从特征选取、模型方法、训练数据集、评价指标值等方面进行了对比分析。最后对虚假评论识别领域的有标注语料规模限制等未来研究方向进行了探讨。     

A study on tolerable waiting time: how long are Web users willing to wait?

Web users often face a long waiting time for downloading Web pages. Although various technologies and techniques have been implemented to alleviate the situation and to comfort the impatient users, little research has been done to assess what constitutes an acceptable and tolerable waiting time for Web users. This research reviews the literature on computer response time and users' waiting time for download of Web pages, and assesses Web users' tolerable waiting time in information retrieval. It addresses the following questions through an experimental study: What is the effect of feedback on users' tolerable waiting time? How long are users willing to wait for a Web page to be downloaded before abandoning it? The results from this study suggest that the presence of feedback prolongs Web users' tolerable waiting time and the tolerable waiting time for information retrieval is approximately 2 s.     

Exploiting Consumer Reviews for Product Feature Ranking

Web 2.0 technology leads Web users to publish a large number of consumer reviews about products and services on various websites.Major product features extracted from consumer reviews may let product providers find what features are mostly cared by consumers,and also may help potential consumers to make purchasing decisions.In this work,we propose a linear regression with rules-based approach to ranking product features according to their importance.Empirical experiments show our approach is effective and promising.We also demonstrate two applications using our proposed approach.The first application decomposes overall ratings of products into product feature ratings.And the second application seeks to generate consumer surveys automatically.     

基于top-k显露模式的商品对比评论分析

随着电子商务的发展,许多购物网站都提供商品评论作为用户购物的决策参考。由于商品评论具有海量、冗余、不规范的特点,用户难以在短时间内浏览所有商品评论,更难以基于评论内容发现商品对比特征。对此,设计了top-k显露模式挖掘算法,并将此算法应用于商品评论对比分析,实现了用户购物决策支持系统——ReviewScope。ReviewScope能够从不同商品的评论中发现特定商品的对比评论,并以此作为购物决策可视化地提供给用户。基于京东商城真实商品评论数据的实验结果表明ReviewScope具有有效、灵活、用户友好的特点。     

What we know about dyslexia and Web accessibility: a research review

Compared to the online interaction behavior of other users, little is known about the difficulties dyslexic Web users encounter online. This paper reviews existing literature at the intersection of dyslexia and accessibility research to determine what useful knowledge exists regarding this important and relatively large group of users. This review uncovers that, although there are few published usability tests with dyslexic users, there is a considerable body of knowledge on dyslexia as well as many design guidelines for authoring dyslexic-accessible interfaces. Through a comparison of existing accessibility guidelines for dyslexic and non-dyslexic users and discussion of the plain language movement, it is argued that dyslexic-accessible practices may redress difficulties encountered by all Internet users. This conclusion suggests that usability testing yielding a clearer profile of the dyslexic user would further inform the practice of universal design, but also that enough knowledge is already available to allow doing more to increase accessibility for dyslexic Internet users.     

W3C user agent accessibility guidelines 1.0 for graphical Web browsers

Web browsers and multimedia players play a critical role in making Web content accessible to people with disabilities. Access to Web content requires that Web browsers provide users with final control over the styling of rendered content, the type of content rendered and the execution of automated behaviors. The features available in Web browsers determine the extent to which users can orient themselves and navigate the structure of Web resources. The World Wide Web Consortium (W3C) User Agent Guidelines are part of the W3C Web Accessibility Initiative, the guidelines provide a comprehensive resource to Web browser and multimedia developers on the features needed to render Web content more accessibly to people with disabilities. UAAG 1.0 was developed over a period of four years and included extensive reviews to demonstrate that the proposed requirements can be implemented.     

A survey on solutions and main free tools for privacy enhancing Web communications

Concern for privacy when users are surfing on the Web has increased recently. Nowadays, many users are aware that when they are accessing Web sites, these Web sites can track them and create profiles on the elements they access, the advertisements they see, the different links they visit, from which Web sites they come from and to which sites they exit, and so on. In order to maintain user privacy, several techniques, methods and solutions have appeared. In this paper we present an analysis of both these solutions and the main tools that are freely distributed or can be used freely and that implement some of these techniques and methods to preserve privacy when users and surfing on the Internet. This work, unlike previous reviews, shows in a comprehensive way, all the different risks when a user navigates on the Web, the different solutions proposed that finally have being implemented and being used to achieve Web privacy goal. Thus, users can decide which tools to use when they want navigate privately and what kind of risks they are assuming.     

Who and what links to the Internet Archive

The Internet Archive’s (IA) Wayback Machine is the largest and oldest public Web archive and has become a significant repository of our recent history and cultural heritage. Despite its importance, there has been little research about how it is discovered and used. Based on Web access logs, we analyze what users are looking for, why they come to IA, where they come from, and how pages link to IA. We find that users request English pages the most, followed by the European languages. Most human users come to Web archives because they do not find the requested pages on the live Web. About 65 % of the requested archived pages no longer exist on the live Web. We find that more than 82 % of human sessions connect to the Wayback Machine via referrals from other Web sites, while only 15 % of robots have referrers. Most of the links (86 %) from Websites are to individual archived pages at specific points in time, and of those 83 % no longer exist on the live Web. Finally, we find that users who come from search engines browse more pages than users who come from external Web sites.