中文网络评论的IT产品特征挖掘及情感倾向分析

为探索中文客户评论中的IT产品特征及相关情感倾向的挖掘,帮助IT生产商和服务商提高改进产品和服务质量,提高竞争力。该文将采用情感分析技术,提出基于客户感知价值的产品特征挖掘算法,实现对于评论中IT产品特征及其情感倾向的语义分析、动态提取和综合信息挖掘;并根据用户的关注权重将产品特征和情感倾向进行排列。采用从互联网下载的真实IT产品评论语料中进行实验,初步验证了该方法的有效性。     

Variability extraction and modeling for product variants

Fast-changing hardware and software technologies in addition to larger and more specialized customer bases demand software tailored to meet very diverse requirements. Software development approaches that aim at capturing this diversity on a single consolidated platform often require large upfront investments, e.g., time or budget. Alternatively, companies resort to developing one variant of a software product at a time by reusing as much as possible from already-existing product variants. However, identifying and extracting the parts to reuse is an error-prone and inefficient task compounded by the typically large number of product variants. Hence, more disciplined and systematic approaches are needed to cope with the complexity of developing and maintaining sets of product variants. Such approaches require detailed information about the product variants, the features they provide and their relations. In this paper, we present an approach to extract such variability information from product variants. It identifies traces from features and feature interactions to their implementation artifacts, and computes their dependencies. This work can be useful in many scenarios ranging from ad hoc development approaches such as clone-and-own to systematic reuse approaches such as software product lines. We applied our variability extraction approach to six case studies and provide a detailed evaluation. The results show that the extracted variability information is consistent with the variability in our six case study systems given by their variability models and available product variants.     

Understanding what concerns consumers: a semantic approach to product feature extraction from consumer reviews

The Web has become an excellent source for gathering consumer opinions (more specifically, consumer reviews) about products. Consumer reviews are essential for retailers and product manufacturers to understand the general responses of customers to their products and improve their marketing campaigns or products accordingly. In addition, consumer reviews enable retailers to recognize the specific preferences of each customer, which facilitates effective marketing decisions. As the number of consumer reviews expands, it is essential and desirable to develop an efficient and effective sentiment analysis technique that is capable of extracting product features stated in consumer reviews (i.e., product feature extraction) and determining the sentiments (positive or negative semantic orientations) of consumers for these product features (i.e., opinion orientation identification). Product feature extraction is critical to sentiment analysis, because its effectiveness significantly affects the performance of opinion orientation identification, as well as the ultimate effectiveness of sentiment analysis. Therefore, this study concentrates on product feature extraction from consumer reviews. Specifically, we propose a semantic-based product feature extraction (SPE) technique that exploits a list of positive and negative adjectives defined in the General Inquirer to recognize opinion words semantically and subsequently extract product features expressed in consumer reviews. Using a prevalent product feature extraction technique and the SPE-GI technique (a variant of SPE) as performance benchmarks, our empirical evaluation shows that the proposed SPE technique outperforms both benchmark techniques.     

Incorporating customer personalization preferences in open product architecture design

Open product architecture is a key enabler for product personalization, as it allows the integration of personalized modules in a product architecture to satisfy individual customer needs and preference. A critical challenge for integrating personalized modules into a product architecture is determining the optimal assembly architecture when considering market expectations and manufacturing constraints. In this paper, an optimization method is proposed for determining the personalized product design architecture that incorporates individual customer preferences. First, a decision hierarchy is presented to describe the integrated design decisions of the product architecture, including product variety determination, module variant selection, and personalized module configuration. Next, a profit model is formulated as an overall performance metric that incorporates customer preferences and manufacturing cost. The systematic patterns and randomness of diverse customer preferences are modeled by combining conjoint analysis and market segmentation with a multivariate normal mixture model. Individual customer product utilities in the target market and their product purchase intent probability are estimated through Monte-Carlo simulation, which is incorporated into the profit calculation. Manufacturing limitations on processes and materials are included as they influence manufacturer’s planning on candidate module variants and production strategies of personalized modules. These models are used to determine a product family architecture that maximizes profit by optimally determining its offering of product variants, module combinations, and personalized module configuration through a genetic algorithm. The proposed method is demonstrated by a personalized bicycle architecture design example.     

Design of a feature-based order acceptance and scheduling module in an ERP system

Customer-oriented manufacturing competes on timely responses to customer requirements, and precise scheduling control for delivery. This challenge demands engineering design and production planning to be fully integrated via advanced enterprise resource planning (ERP) systems. This paper proposes a generic feature association method and a detailed framework that can unify product and process models in order to satisfy customer orders with small batch sizes and high variations. A conceptual solution is introduced by integrating two traditionally separate feature domains: design configuration features and manufacturing process features. To achieve the proposed method, a customer feature class is suggested for the characterization of customers’ profiles related to the manufacturer. An instantiated customer feature object functionally tracks each customer’s selection of product configurations related to its requirements, specific orders, and production schedules with dynamic associations to the live manufacturing capacity. With the new associative integration method, a preliminary order acceptance system (OAS) prototype system has been implemented within an ERP order management system and its conceptual structure model is demonstrated within a multi-facet feature framework.     

Dynamic modelling of customer preferences for product design using DENFIS and opinion mining

Previous studies mainly employed customer surveys to collect survey data for understanding customer preferences on products and developing customer preference models. In reality, customer preferences on products could change over time. Thus, the time series data of customer preferences under different time periods should be collected for the modelling of customer preferences. However, it is difficult to obtain the time series data based on customer surveys because of long survey time and substantial resources involved. In recent years, a large number of online customer reviews of products can be found on various websites, from which the time series data of customer preferences can be extracted easily. Some previous studies have attempted to analyse customer preferences on products based on online customer reviews. However, two issues were not addressed in previous studies which are the fuzziness of the sentiment expressed by customers existing in online reviews and the modelling of customer preferences based on the time series data obtained from online reviews. In this paper, a new methodology for dynamic modelling of customer preferences based on online customer reviews is proposed to address the two issues which mainly involves opinion mining and dynamic evolving neural-fuzzy inference system (DENFIS). Opinion mining is adopted to analyze online reviews and perform sentiment analysis on the reviews under different time periods. With the mined time series data and the product attribute settings of reviewed products, a DENFIS approach is introduced to perform the dynamic modelling of customer preferences. A case study is used to illustrate the proposed methodology. The results of validation tests indicate that the proposed DENFIS approach outperforms various adaptive neuro-fuzzy inference system (ANFIS) approaches in the dynamic modelling of customer preferences in terms of the mean relative error and variance of errors. In addition, the proposed DENFIS approach can provide both crisp and fuzzy outputs that cannot be realized by using existing ANFIS and conventional DENFIS approaches.     

中文网络评论的产品特征提取及情感倾向判定

随着互联网和电子商务的发展,用户在购买或使用商品之后会在网络站点上发表对产品的评论,大量的产品评论中所包含的丰富信息,可以为生产厂商和用户提供重要的决策依据。基于文本的语义和语言分析,提出了从产品评论中提取用户关注的产品特征的方法,并根据用户的关注程度对产品特征进行排序;同时,根据观点词的极性值判定用户对产品特征的情感倾向以及情感倾向强度。本研究采用从互联网上获得的针对笔记本电脑的产品评论作为实验对象,实验结果初步证明该方法具有良好的准确率和召回率。     

Deep sentiment analysis: Mining the causality between personality-value-attitude for analyzing business ads in social media

IT vendors routinely use social media such as YouTube not only to disseminate their IT product information, but also to acquire customer input efficiently as part of their market research strategies. Customer responses that appear in social media, however, are typically unstructured; thus, a fairly large data set is needed for meaningful analysis. Although identifying customers’ value structures and attitudes may be useful for developing targeted or niche markets, the unstructured and volume-heavy nature of customer data prohibits efficient and economical extraction of such information. Automatic extraction of customer information would be valuable in determining value structure and strength. This paper proposes an intelligent method of estimating causality between user profiles, value structures, and attitudes based on the replies and published content managed by open social network systems such as YouTube. To show the feasibility of the idea proposed in this paper, information richness and agility are used as underlying concepts to create performance measures based on media/information richness theory. The resulting deep sentiment analysis proves to be superior to legacy sentiment analysis tools for estimation of causality among the focal parameters.     

Sentiment analysis from Customer-generated online videos on product review using topic modeling and Multi-attention BLSTM

With the popularity of social websites and mobile applications including Instagram, YouTube, TikTok, etc., online videos shared by customers presenting their thoughts and reviews on products are posted daily in increasing numbers. Such online videos containing Voice of Customer (VOC) are precious for product designers or managers to capture customer sentiment and understand customer preference. For this purpose, we propose a novel method for analyzing customer sentiment from online videos on product review. Firstly, latent Dirichlet allocation (LDA) modeling is applied to identify the topics from the online videos after data preprocessing. Then sentiment polarity corresponding to each topic of each speaker in videos can be identified using our newly designed multi-attention bi-directional LSTM (BLSTM(MA)), which can better mine complex relationships among a speaker’s sentiments on different topics. This paper is of great practical value for company managers and researchers to better understand a large number of customer opinions on specific products. To explain the application of this method and prove its effectiveness, two cases respectively on smartphones and several published datasets are developed finally.     

PRONTO: An ontology for comprehensive and consistent representation of product information

Nowadays, it is quite common for collaborating organizations (or even different areas within a company) to develop and maintain their own product model. This situation leads to information duplication and its associated problems. Besides, traditional product models do not properly handle the high number of variants managed in today competitive markets. In addition, there is a need for an integrated product model to be shared by all the organizations participating in global supply chains (SCs) or all the areas within a company. One way to reach an intelligent integration among product models is by means of an ontology. PRoduct ONTOlogy (PRONTO) is an ontology for the product modeling domain, able to efficiently handle product variants. It defines and integrates two hierarchies to represent product information: the abstraction hierarchy (AH) and the structural one (SH). This contribution presents a ConceptBase formal specification of PRONTO that focuses on the structural hierarchy of products. This hierarchy is a tool to handle product information associated with the multiple available recipes or processes to manufacture a particular product or a set of similar products. The formal specification presented in the paper also includes mechanisms to infer structural information from the explicit knowledge represented at each of the AH levels: Family, VariantSet and Product. This proposal efficiently handles a great number of variants and allows representing product information with distinct granularity degrees, which is a requirement for planning activities taking place at different time horizons. PRONTO easily manages crucial features that should be taken into account in a product representation, such as the efficient handling of product families and variants concepts, composition and decomposition structures and the possibility of specifying constraints. To demonstrate the semantic expressiveness of the proposed ontology a food industry related case-study is addressed and discussed in detail.     

基于性格情绪特征的改进主题情感模型

近年来,以微博为代表的社交媒体在情感分析中备受关注。然而,绝大多数现有的主题情感模型并没有充分考虑到用户性格特征,导致情感分析结果难尽人意。故该文在现有的JST模型基础上进行改进,提出一种基于时间的性格建模方法,将用户性格特征纳入主题情感模型中;鉴于微博数据包含大量的表情符号之类的特有信息,为了充分利用表情符号来提升微博情感识别性能,该文将情感符号融入JST模型中,进而提出了一种改进的主题情感联合模型UC-JST(Joint Sentiment/Topic Model Based on User Character)。通过在真实的新浪微博数据集上进行实验,结果表明UC-JST情感分类效果优于JST、TUS-LDA、JUST、TSMMF四种典型的无监督情感分类方法。     

Performance variability in software product lines: proposing theories from a case study

In the software product line research, product variants typically differ by their functionality and quality attributes are not purposefully varied. The goal is to study purposeful performance variability in software product lines, in particular, the motivation to vary performance, and the strategy for realizing performance variability in the product line architecture. The research method was a theory-building case study that was augmented with a systematic literature review. The case was a mobile network base station product line with capacity variability. The data collection, analysis and theorizing were conducted in several stages: the initial case study results were augmented with accounts from the literature. We constructed three theoretical models to explain and characterize performance variability in software product lines: the models aim to be generalizable beyond the single case. The results describe capacity variability in a base station product line. Thereafter, theoretical models of performance variability in software product lines in general are proposed. Performance variability is motivated by customer needs and characteristics, by trade-offs and by varying operating environment constraints. Performance variability can be realized by hardware or software means; moreover, the software can either realize performance differences in an emergent way through impacts from other variability or by utilizing purposeful varying design tactics. The results point out two differences compared with the prevailing literature. Firstly, when the customer needs and characteristics enable price differentiation, performance may be varied even with no trade-offs or production cost differences involved. Secondly, due to the dominance of feature modeling, the literature focuses on the impact management realization. However, performance variability can be realized through purposeful design tactics to downgrade the available software resources and by having more efficient hardware.     

基于并行混合网络与注意力机制的文本情感分析模型

现有基于深度学习和神经网络的文本情感分析模型通常存在文本特征提取不全面,且未考虑关键信息对文本情感倾向的影响等问题。基于并行混合网络与双路注意力机制,提出一种改进的文本情感分析模型。根据不同神经网络的特点分别采用GloVe和Word2vec两种词向量训练工具将文本向量化,得到更丰富的文本信息。将两种不同的词向量并行输入由双向门控循环单元与卷积神经网络构建的并行混合网络,同时提取上下文全局特征与局部特征,提高模型的特征提取能力。使用双路注意力机制分别对全局特征和局部特征中的关键信息进行加强处理及特征融合,增强模型识别关键信息的能力。将融合后的整个文本特征输入全连接层,实现最终的情感极性分类。在IMDb和SST-2公开数据集上的实验结果表明,该模型的分类准确率分别达到91.73%和91.16%,相比于同类文本情感分析模型有不同程度的提升,从而证明了双路注意力机制可以更全面地捕获文本中的关键信息,提高文本情感分类效果。     

SumView: A Web-based engine for summarizing product reviews and customer opinions

In this paper, we develop SumView, a Web-based review summarization system, to automatically extract the most representative expressions and customer opinions in the reviews on various product features. Different from existing review analysis which makes more efforts on sentiment classification and opinion mining, our system mainly focuses on summarization, i.e., delivering the majority of information contained in the review documents by selecting the most representative review sentences for each extracted product feature. Comprehensive case studies and experiments demonstrate the effectiveness of our system, and the user study shows users’ satisfaction.     

基于RoBERTa-WWM的大学生论坛情感分析模型

大学生论坛语句具有篇幅短、口语化、多流行用语等特点,传统的情感分析模型难以对其进行精准的语义特征表示,并且未充分关注语句的局部特征与上下文语境。提出一种基于RoBERTa-WWM的大学生情感分析模型。通过RoBERTa-WWM模型将论坛文本语句转化为语义特征表示,并将其输入到文本卷积神经网络中,以提取语句的局部语义特征,同时利用双向门控循环单元网络对局部语义特征进行双向处理,获得全面的上下文语义信息。在此基础上,通过Softmax分类器计算语句在情感标签中的概率向量,选择最大值表示的情感标签作为最终输出的情感标签。实验结果表明,相比RoBERTa-WWM、EK-INIT-CNN、BERT等模型,该模型在大学生论坛与NLPCC2014数据集上具有较优的分类性能,并且在大学生论坛数据集上宏平均精准率、宏平均召回率、宏平均F1值和微平均F1值分别为89.43%、90.43%、90.12%和92.48%。     

基于非完备信息系统的评价对象情感聚类

该文利用领域本体对产品评论文本中的评价对象进行抽取和整合,在此基础上,建立产品性能的非完备信息系统,将特征的情感倾向寓于特征的权重计算之中。对非完备信息系统,给出了基于差别矩阵的启发式特征约简方法,通过特征降维处理,达到了减少特征的冗余度和数据稀疏性的目的。对降维后的非完备信息系统采用K-Means聚类算法,实现了评价对象情感聚类。为了验证该文提出方法的有效性,在真实汽车评论文本数据上进行实验, 实验结果表明,在对特征进行一定程度的降维后,仍表现出较好的聚类效果。     

Customized products recommendation based on probabilistic relevance model

Product customization is attracting more attentions in industry as a viable strategy to better meet customer requirements and gain more profit. However the vast number of product variants in product customization process often makes it difficult for consumers to make purchase decisions, a phenomenon referred to as information overload. In this paper we take a two-prong approach to tackle the issue of information overload in customized products recommendation. Basically, the method answers two questions, namely, which products to recommend and in what order to present the recommendations. Firstly, a probability relevance model is deployed to calculate the probability of relevance for each end product. Then a probability ranking principle is exploited to present the recommendations. The approach also takes customer flexibility into consideration and thus mitigates the effect of inconsistent specifications from customers. It does not require any prior knowledge about an active customer’s preference and can accommodate the new customers challenge facing by recommendation approaches. Analytical results show that the method is optimal in terms of customer’s utility and product recommendation efficiency. Numerical experiments are also conducted to test the presented approach.     

Multi-level feature optimization and multimodal contextual fusion for sentiment analysis and emotion classification

The availability of the humongous amount of multimodal content on the internet, the multimodal sentiment classification, and emotion detection has become the most researched topic. The feature selection, context extraction, and multi-modal fusion are the most important challenges in multimodal sentiment classification and affective computing. To address these challenges this paper presents multilevel feature optimization and multimodal contextual fusion technique. The evolutionary computing based feature selection models extract a subset of features from multiple modalities. The contextual information between the neighboring utterances is extracted using bidirectional long-short-term-memory at multiple levels. Initially, bimodal fusion is performed by fusing a combination of two unimodal modalities at a time and finally, trimodal fusion is performed by fusing all three modalities. The result of the proposed method is demonstrated using two publically available datasets such as CMU-MOSI for sentiment classification and IEMOCAP for affective computing. Incorporating a subset of features and contextual information, the proposed model obtains better classification accuracy than the two standard baselines by over 3% and 6% in sentiment and emotion classification, respectively.     

Testing variability-intensive systems using automated analysis: an application to Android

Software product lines are used to develop a set of software products that, while being different, share a common set of features. Feature models are used as a compact representation of all the products (e.g., possible configurations) of the product line. The number of products that a feature model encodes may grow exponentially with the number of features. This increases the cost of testing the products within a product line. Some proposals deal with this problem by reducing the testing space using different techniques. However, a daunting challenge is to explore how the cost and value of test cases can be modeled and optimized in order to have lower-cost testing processes. In this paper, we present TESting vAriAbiLity Intensive Systems (TESALIA), an approach that uses automated analysis of feature models to optimize the testing of variability-intensive systems. We model test value and cost as feature attributes, and then we use a constraint satisfaction solver to prune, prioritize and package product line tests complementing prior work in the software product line testing literature. A prototype implementation of TESALIA is used for validation in an Android example showing the benefits of maximizing the mobile market share (the value function) while meeting a budgetary constraint.     

基于神经网络模型的产品属性情感分析

针对基于词向量的神经网络模型在产品属性情感分析中效果不佳的问题,提出一种集成离散特征和词向量特征的开关递归神经网络模型。首先,通过直接循环图为语句建模,采用开关递归神经网络模型完成产品属性情感分析任务;然后,在开关递归神经网络模型中集成离散特征和词向量特征;最后,分别在流水线、联合、折叠三种任务模型中完成属性提取和情感分析任务。以宏观F1分数作为评估指标,在SemEval-2014的笔记本电脑和餐馆评论数据集上做实验。开关递归神经网络模型的F1分数为：48.21%和62.19%,超过普通递归神经网络模型近1.5个百分点,因而开关递归神经网络能够有效捕获复杂特征,提升产品属性情感分析的效果。而集成离散特征和词向量特征的神经网络模型的F1分数为：49.26%和63.31%,均超过基线结果0.5到1个百分点,表明离散特征和词向量特征互相促进,另一方面,也表明仅仅基于词向量的神经网络模型仍有提升空间。三种任务模型中,流水线模型的F1分数最高,表明应将属性提取和情感分析任务分开完成。