- About this Journal ·
- Abstracting and Indexing ·
- Aims and Scope ·
- Annual Issues ·
- Article Processing Charges ·
- Author Guidelines ·
- Bibliographic Information ·
- Citations to this Journal ·
- Contact Information ·
- Editorial Board ·
- Editorial Workflow ·
- Free eTOC Alerts ·
- Publication Ethics ·
- Recently Accepted Articles ·
- Reviewers Acknowledgment ·
- Submit a Manuscript ·
- Subscription Information ·
- Table of Contents
Advances in Mechanical Engineering
Volume 2013 (2013), Article ID 536820, 10 pages
Customer Preference-Based Information Retrieval to Build Module Concepts
1School of Mechanical Engineering, Hebei University of Technology, Tianjin 300130, China
2School of Computer Science and Engineering, Hebei University of Technology, Tianjin 300130, China
3EaglePicher Medical Power, Plano, TX 75025, USA
Received 11 April 2013; Accepted 3 July 2013
Academic Editor: Yu-Shen Liu
Copyright © 2013 Dongxing Cao et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Preference is viewed as an outer feeling of a product, also as a reflection of human's inner thought. It dominates the designers’ decisions and affects our purchase intention. In the paper, a model of preference elicitation from customers is proposed to build module concepts. Firstly, the attributes of customer preference are classified in a hierarchy and make the surveys to build customer preference concepts. Secondly, the documents or catalogs of design requirements, perhaps containing some textual description and geometric data, are normalized by using semantic expressions. Some semantic rules are developed to describe low-level features of customer preference to construct a knowledge base of customer preference. Thirdly, designers’ needs are used to map customer preference for generating module concepts. Finally, an empirical study of the stapler is surveyed to illustrate the validity of module concept generation.
Why do the customers prefer to buy a kind of products and not others? Is that their brands, shapes, or reliabilities? How to reduce the risk of product development from evaluating the choices customers make? Of course, if customers could exactly indicate their preference, product development would be fairly risk-free [1, 2]. However, the fact is the opposite, and the uncertainty still appears at the earlier design stage of product development. The difference maybe appears in between designers and customers as the developed product is not what the customer wants [3, 4]. Therefore, in order to reduce risks and fasten the lead time to market, product development should take into consideration rapid responses in customers’ voices . In general, design can be viewed as an iterative process, in which customer preference dominates the final result of the conceptual design of a product .
At the early stage of product concept generation, customer preference has a direct impact on the number of iterative design, scheme evaluation, and cost. It also will affect the final design decision. Even when reliable customer information is elicited, it is still a great challenge to efficiently implement conceptual design . Customer satisfaction degree can be also described by a conventional fuzzy set, in which fuzzy rules are employed to model the knowledge to evaluate design candidates . Fuzzy reasoning is carried out to obtain the importance of customer needs and design metrics. In the preliminary design stage, the specification of design is usually incomplete. The compromise between designers and customers can be obtained by their negotiations to make a concession with each other.
In this paper, we propose a customer preference model for new product development by automatically mining the text information of customer preference online product reviews . For fast developing products, customer’s perceptions and product module concepts are constantly changing depending on customer preference. An automatic analysis approach is fast and simple, and it can remain abreast of changes in the market. In this research, firstly, we review the related literatures at the customer preference. Secondly, we present the proposed approach, and the attributes of customer preference are classified in a hierarchy. Next, the process of ontology preference modeling is described to realize preference semantic extraction. Then, an empirical study is applied to generate module concepts.
2. Related Work
There are several methods for surveying customers’ preference. They mainly include conjoint analysis, contingent valuation, direct value surveying, house of quality, and utility analysis [5, 10], in which conjoint analysis requires the survey subject to jointly consider several product attributes. These hypothetical product attributes are defined as a designed experiment, intended to obtain the most accurate information from the least amount of input. Contingent valuation involves directly asking survey respondents about their willingness to pay for a particular product feature. Direct value survey is utilized in value engineering and is a method that integrates features of conjoint analysis, contingent valuation, utility analysis, and prospect theory. House of Quality (HOQ) is a well-known method for eliciting, organizing, and communicating the “voice of the customer” to the design process . A matrix structure is employed. On the vertical axis, the voice of the customer is expressed in a list of customer’s desires. These might be entries such as “low cost,” “light weight,” or “environmentally friendly.” The horizontal axis contains design variables that the engineer can directly control to satisfy the stated customer desires. Entries within the matrix and in the “roof” indicate interrelationships between each entry. Utility analysis is an axiomatically based approach to decision making, and it is particularly useful for dealing with decision making under uncertainty. Hauge and Stauffer  developed the taxonomy of customer requirements as an initial concept graph structure for questionnaires used for an expert system. The approach was aimed at further elicitation of knowledge from customers.
Since our objective mainly focuses on obtaining module concepts and performing the function of a product, it is also imperative to capture the customer voices. In this work, we discuss how to generate preference knowledge base by processing online product reviews and adaptive text extraction. The objective makes efforts to bridge the gap between low level attributes and high level attributes as shown in Figure 1 .
In this research, we carry out the survey by articulating two distinct problems: needs and attribute specifications. First, we differentiate between product attributes and user needs by proposing a method that explicitly addresses needs identification. Second, concept generation requires more than attribute classifications. A single product attribute is specified by using multiple properties and levels. However, our approach is used to cluster and manage categorical data. In the context of words and phrases in reviews, we might treat every word as a node in the graph. Edges would represent the relationship between words. Edges are weighted based upon the number of reviews in which the two words appear together .
3. The Proposed Approach
Most of the existing information is disorderly arranged and unsystematic. It needs product designers’ analysis and extraction in order to provide useful information . At the beginning of design, the original content of preference from customers, such as from survey reports, transaction data, and customer dialogs, should be filtered. A normal design information text or document is used to extract preference information after transformation . Automatically extracting semantics from the normalized document requires recognizing the syntactic structure as well as the semantic meaning of the text. Linguistic knowledge and domain knowledge are needed to fulfill preference semantic extraction.
Figure 2 gives the infrastructure of a prototype for design preference information extraction. First of all, the original material documents from customers, such as user requirements, survey reports, and transaction data, and are acquired and then transformed into design information texts or documents in which some unstructured information should adopt an effective approach. Second, some terms and concepts can be extracted from preference semantic structures, such as noun phrases, verb phrases, adverb phrases, and adjective phrases. These can be represented by using an ontology-based design semantic analysis and information extraction. The taxonomy of preference is classified to acquire the relationships between two concepts further. The specific thesauri or lexicons are built to capture preference concepts from the knowledge base of preference. The ontology expression and preference semantics of extraction process are described, and they are based on a shallow NLP algorithm and the domain ontology. Next, the preference ontology concepts are acquired after the extraction process. The extraction algorithm and the preference metrics are described and are used for preference ontology modeling. Finally, an empirical study for design preference extraction is introduced to build module concepts.
4. Customer Preference Analysis
4.1. Attribute of the Preference
The enterprises aim to set up a great reputation for their product on customers’ minds. They often inquire about customers in order to find out the needs that are not met by existing products and assess demand degrees for a new product where no product currently exists. Then, they define the product in terms of attributes of preference and develop the product to settle for the market demands. After finishing the analysis of the need preference, product designers can work towards concept generation in order to customize product configurations. However, the preference cannot be viewed as equivalent to demand. Their categories are different. Preference has subjectivity and is related to the behavior of customer’s feelings, whereas demand is more objective and mainly depends on other factors, such as availability, familiarity, word of mouth, advertising, and store or shelf location.
Thus, customer preference (CP) includes instinct factors (IF), aesthetic sensibilities (AS), emotional factors (EF), faiths (Fa), appreciation abilities (AA), and time span (TS). There, the instinct factors stand for product characteristics which include functions, performances, reliabilities, cost, and lifecycle. Aesthetic sensibilities include shape, color, operation, decoration, and feel. Faiths include religion, ethnologic culture, and race prejudice. Emotional factors affect the intention of purchase, decide the customers’ mood, and include good temper, bad temper, indifferent, inpatient mood, and irritable mood. Appreciation abilities mean the customers’ cognition and their contents include knowledge background, educational degree, and habitat. Time span means how many years it lasts, and it is short time or long time. Attribute of the preference is shown in Figure 3.
Customer preference is of certain relativity and is not absolute . It is changeable relative to time span, scene, and attribute. The time span, which depends on the category of the product, is uncertain. Sometimes it is long, sometimes short. The scene, which depends on customer behavior, is associated with different cultures and geographies . In addition, customer preference may change, when the value of some attributes is changed, such as function, shape, and cost.
Therefore, customer preference can be formally represented as follows: where IF = (function, performance, reliabilities, cost, and lifecycle), AS = (shape, color, operation, decoration, and feel), EF = (good temper, bad temper, indifference, inpatient mood, and irritable mood), Fa = (religion, ethnologic culture, and race prejudice), AA = (knowledge background, educational degree, and habitat), and TS = (year, short time, and long time).
4.2. Survey of the Preference
A best-selling product is definitely based on a favorable customer preference. First, the main factors from customer perspectives should be identified and the domain knowledge of the product should be collected in a professional survey before the product is launched. Second, a survey activity is conducted to determine the customers’ needs and desires before putting the new product on the market. This survey can be analyzed by using a software tool to determine the specific customer preference. Based on this consideration, a measure about the acceptance of potential customers can be taken and market simulations can become feasible. Therefore, customer preference can support demand analysis, conceptual design, and embodiment design, and at the same time, they are related to direct and indirect surveys and experiment results, as shown in Figure 4.
Enterprises routinely ask their customer preference questions, that is, direct survey, and in this way, they can provide better service to their customers or improve customers’ satisfaction degree of products . Surveys are sometimes described as informal conversations between product designers and customers. A number of measures can be taken to conduct the survey . A very common method is to ask customers questions about their preference for particular product function, feature, shape, cost, or even service quality. A scale is commonly used in survey questions to elicit preference or evaluations. The value of the labels perhaps has a prejudice against the selected results which depend on customers’ personal desires.
Indirect survey is another traditional approach to generate different concepts of a design and to conduct experiments with customers to capture preference . By using a software tool as the customer service platform, it is much easier to run experiments on website. And most of them always ran such experiments and showed a raise in the browsing rate through clicking and determining whether a new design increased sales in a few days. If a product is advertised online, we will discover in a few hours whether experiment results or ad click rates increase, and the transaction data reveals copurchasing patterns for customer services. Some popular websites certainly have a high ad click rate. When customers go shopping, the transaction data can reveal their preference for a particular product and enable results targeted to specific buyer groups or buyer categories. How easy it is to identify customer preference depends on the context online and on the customer's willingness to buy a product. However, these text descriptions are disorganized but do have a great deal of information. It is necessary to extract customer preference by employing some effective methods, such as AHP , statistic method , and decision algorithm . Therefore, customer preference has the multidimensional properties, such as price, features, quality, performance, brand, distribution channel, safety, and usability.
5. Ontology Preference Modeling
5.1. Expression of Preference Ontology
Ontology is a formal, explicit specification of a shared conceptualization , where conceptualization refers to an intended model of the world’s phenomena identified by its concepts and relation. Explicit means that the concepts and relations are explicitly defined, while formal means that it can be communicated across people and computer. Therefore, ontology defines a set of representational terms we call concepts. They should represent the hierarchical correlations among concepts . On the other hand, the taxonomy is only reviewed as concept classification in the hierarchy. It simply links concepts by domain-independent relationships.
Ontology concepts have multiple parents and form the complex relations of inheritances. They share the genetic attributes. At present, considering ontology modeling for customer preference, there are two main problems: one is the extraction of the semantic concepts from the preference words, and the other is the document indexing from customers’ requirements. As for the first problem, the key issue is to identify appropriate preference concepts and build preference lexicon based on design documents, at the same time, indexing the preference terms from customer documents, in which the precision problem of extraction is about the semantic expression employed in customer requests. A hierarchical analysis process has been used to aggregate preference in a group using a pair-wise approach . However, a significant assumption is assumed to be equally important. That is, the information is handled equally without any preference given to one group member over another.
Ontology modeling provides an effective approach to indexing terms/concepts which can be used to match with customer requests. However, the taxonomy acquisition of customer preference of different products is of a certain subjective behavior. Their generation is either by brainstorming or by interviewing or dialoging with customers. Under similar circumstances, we can acquire preference ontologies . Figure 5 presents the taxonomy of customer preference ontology, which comes from stapler handbooks or knowledge resources. For example, stapler handbooks often classify engineering components which can be clustered into an ontology model as concepts and taxonomy in the hierarchy. Each component is described in detail, including its attributes such as instinct, aesthetic, emotional, and religious, which can easily be identified and mapped to preference ontologies as well as corresponding relationships.
Customer preference ontology includes concepts, taxonomies, and relationships. Each taxonomical concept is acquired from various engineering knowledge resources. We can adopt terms or phrases to describe the concepts of the taxonomy as well as their relationships with other concepts. For example, magazine belongs to IF taxonomy of the stapler. We can represent this as IF-MAGAZINE, where the prefix of each concept represents the taxonomy which this concept belongs to. Therefore, the relationships are structured between concepts across taxonomies. For example, relationship “has_feature” has a concept “AS-SIVER/AS-BOX LIKE-STAPLER” as shown in Table 1, in which AS-SIVER stands for a color concept in the instinct taxonomy, AS-BOX LIKE-STAPLER represents a shape concept in the aesthetic taxonomy [14, 21]. Table 1 lists customer preference ontological concepts and their relational classification.
5.2. Knowledge Base of Preference Ontology
Generally speaking, lexical terms are words or phrases corresponding preference concepts in the documents. They are used to map the concepts of different texts. Therefore, word morphs, abbreviations, acronyms, and synonyms of the word/phrase are lexical terms and share the same concepts with the original lexical terms . Also, some noun phrases, verb phrases, adverb phrases, and prepositional phrases can be extracted as preference terms. The morphs of original lexical terms can easily and automatically be obtained by WordNet (http://wordnet.princeton.edu/) , whereas other terms can be acquired manually because WordNet is a general lexical resource but not a professional preference lexicon. We aim to extract implicit customer preference based on the domain knowledge. As the existing case studies are almost special products, the extracted texts have a certain limitation. If a preference lexicon is built, it can be used for improving the preference evaluation possibilities. However, it is not easy to model and extract the semantic information of implicit customer preference from design texts which are embedded into natural language. In order to identify linguistic forms of customer preference, we build a preference lexicon to support automatic indexing. Logically speaking, such preference information is implicit within engineering design texts, but it could be difficult to extract from unstructured documents. In order to overcome this difficulty, we build the preference semantic model and its mapping into the ontology concepts. We identify linguistic forms of preference, produce a specific preference lexicon, develop customer preference ontology concepts, and generate design alternatives. A preference lexicon can show what the customers want. Figure 6 presents a common preference lexicon for the stapler, which looks like six planets round the preference lexicon. Each planet has its preference terms and can be decomposed further in the hierarchy. Here, the arrowheads describe the different preference terms which indicate stapler functions, performance, shape, cost, color, and so on.
6. Preference Semantic Extraction
6.1. Expression of Preference Ontology
Semantic ambiguity often occurs in design making when customers do not know the exact expressions or the related concepts they want to pursue though they may have some contextual clues, such as the functional preference of the design and other interacting parts of the product. A preference lexicon is a better way to evaluate customer preference .
Semantic rules are used to link preference terms and concepts together to build the customer preference concepts to aid in searching for design information. In general, there are two types of semantic rules. One is from the combination of preference terms and concepts, in which each term includes a noun, verb, adj, adv, and pron, and each concept is composed of several words. For example, the combination of ‘‘RED” and ‘‘AS-PUPPY DOG-STAPLER” forms a new preference concept ‘‘AS-RED-PUPPY DOG-STAPLER.” The other is the combination of two concepts. For example, ‘‘IF-ANVIL” and ‘‘IF-BASE” constitute a new concept ‘‘IF-ANVIL-BASE.” The concept from semantic synthesis is called Instance Concept (IC) which has a certain entity meaning. Generally speaking, some instance concepts exist in specific relationships, such as is_a, has_function, has_part, and has_material. These relationships are on the basis of forming concept ontology relationships [21, 23].
We use instance concept to construct design semantics through a set of slots and relations . For example, each concept instance has several slots which describe its functions, properties, materials, and relationships. In the process of system work, the documents are scanned to search for instance concept and its specific value. Each concept corresponds to a relative slot value. For example, design object of the stapler has a specific slot ‘‘has_part” which corresponds to the instance concept, ‘‘HAMMER.” It will be scanned and tagged in the process of indexing sentences. Meanwhile, the concept ‘‘HAMMER” has a function slot ‘‘STRIKING MOTION” and its upper cover exists in a material slot ‘‘has_material”; that is, ‘‘UPPER COVER” is made of material “PLASTIC.” In the same way, we can find the stapler function slot ‘‘has_function,” and it has a function “MAGAZINE” and three properties: ‘‘STORING STAPLES, ORIENTATION, and REORIENTATION.” Customer preference ontologies exist in the concept forest which is interlinked by means of conceptual and lexical relations as shown in Figure 7. For example, given words “HAMMER,” “STAPLES,” and “ENVIL,” we can construct a concept tree for terms “HAMMER,” “STAPLES,” and “ENVIL” using “is_a” or “has_part” relationship links based on the hypernymy relationships.
6.2. Preference Concept Extraction and Module Generation
As customers often present their requirements by using natural language; some words, terms, and concepts can be used to represent customer preference. Sometimes different word or term usage can show a big difference in the preference degree . At first, we need to extract some words from documents and dictionaries. These words can be classified as different taxonomies based on semantic description. An analyzer is built to separate these words into different taxonomies. Some independent words will be scaled based on linguist viewpoints. On the other hand, semantic rules are built to compose the concepts, in which rule base is crucial to build concepts. Also according to the viewpoints of linguist, some concepts should be advisably scaled. The scaled words and concepts will be put into preference knowledge base. In addition, we need to classify concept into taxonomies, such as function preference, performance preference, and cost preference.
We can extract different preference concepts from a document based on different semantic structures. It is an effective method to use concept forests to represent the semantic content of a text document; the semantic similarity of two text documents can be determined by comparing their concept forests. Formally, a concept forest (CF) is defined as a graph as follows : where indicates a set of stemmed words, that is, , and stands for a set of edges, that is, , which connect stemmed words with relationships defined in , that is, .
Specifically, an edge is defined as a triplet in the following:
Assuming the two documents and , and their concept forests and , respectively, calculating the semantic similarity of two documents needs to consider the similarities of the term, edge, and relationship sets in their concept forests; therefore, we calculate the semantic similarity of two text documents by simply comparing the similarity of the terms ( and ) in their concept forests, as follows:
Therefore, we can realize document ranking on the basis of the previous semantic similarity. Assuming that module sets , in which stands for term concepts, and according to the degree of similarity measures, term concepts are clustered into a cluster () while we will divide module sets into clusters in the following: where means weight factor of different term concepts. And satisfies the following relationships:
Therefore, we obtain the number of modules () as follows:
7. Empirical Study
A case study on stapler is used to illustrate the proposed method. We collect 200 documents, totally about 1M memories, in which 100 documents come from stapler specification descriptions, and the other 100 documents for customers’ reviews from several familiar websites (http://www.amazon.co.uk/, http://www.ebay.com/, etc.). They delegate most users’ requirements for staplers. We only select text documents excluding pictures, charts, and graphs from websites as they are out of the range. Nine brands are selected from several companies, such as Office Depot, Leitz, Rapid, Stanley, and Swingline.
The objective of module concept generation involves identifying customer requirements and then mapping them into a set of stapler attributes or specifications. Considering this case review, the designer would like to generate concept modules of a new stapler by means of clustering algorithms. A formal or normal document is needed to use for text information extraction. We first extract all keywords and their occurrence frequencies from the document, excluding stop words such as pronouns, common verbs, common nouns, adjectives, and frilly words. These words are of little or no value in determining the document’s semantic content. Some classifications are annotated as part-of-speech (POS) tagging by indexing preference lexicon . On the basis of domain ontology base, the terms/phrases are recognized to build concept forests. By indexing semantic rule base, jointing relationships are established to classify out different preferences (see Section 5.1). Then, module concepts are clustered by an effective algorithm based on (4) and (5). Figure 8 presents clustering chart of stapler module concepts. Different kinds of components will be clustered into the specific functions of stapler. Figure 9 presents module concepts and their components which form different types of models. They are obtained by extracting 100 documents from stapler specification descriptions.
Any preference concepts can be viewed as a part of stapler or one of components within module. Preference terms are indexed and tagged. Similarity measures are used to select preference ontology concepts. After text document extraction, we select nine models which have commercial value staplers as shown in Table 2. They are obtained by indexing 100 documents from customers’ reviews. These results will provide model references to office enterprises for new product development further. Furthermore, enterprises will increase customer preference degree by improving the performance of their existing staplers. At the same time, different sex and age people can still select more suitable staplers for themselves.
In the paper, the customer preference ontology is described and preference design information is extracted to build a preference knowledge base which includes a preference lexicon, domain ontology, and semantic rules. An ontology-based model is given for information retrieval. The concept generation and selection of information are based on customer preference ontology. We have used the preference domain knowledge of the stapler for describing the proposed approach, while the results can be applied to other similar products. However, further research is needed as follows. (i)The large amount of informal design information is steadily increasing on website. These texts are less likely to comply with the formal documental format. It is difficult to extract the ontology concept semantics from these documents. It is worth investigating further in the future.(ii)Information extraction of customer preference is currently based on indexing the sentence semantic rules, in which the preference lexicon and domain knowledge are crucial to achieve information retrieval. However, how to avoid semantic blur and improve the indexing precision? Further work is needed. (iii)At a particular time, customers show a strong liking for certain staplers. But later, they show a change for another stapler. Therefore, how to update dynamically the changes with a fast and simple response to customer preference in the market is needed.
This research is partially sponsored by the National Natural Science Foundation of China (NSFC) under Grant nos. 50775065 and 51275152 and Nature Science Foundation of Hebei Province under Grant nos. E2008000102 and E2013202123.
- E. F. MacDonald, R. Gonzalez, and P. Y. Papalambros, “Preference inconsistency in multidisciplinary design decision making,” Journal of Mechanical Design, vol. 131, no. 3, pp. 0310091–03100913, 2009.
- K. N. Otto and E. K. Antonsson, “Trade-off strategies in engineering design,” Research in Engineering Design, vol. 3, no. 2, pp. 87–103, 1991.
- R. Zhao and W. I. Grosky, “Narrowing the semantic gap: improved text-based web document retrieval using visual features,” IEEE Transactions on Multimedia, vol. 4, no. 2, pp. 189–200, 2002.
- M. Kulok and K. Lewis, “Preference consistency in multiattribute decision making,” in Proceedings of the ASME International Design Engineering Technical Conferences and Computers and Information in Engineering Conference (DETC '05), pp. 291–300, Long Beach, Calif, USA, September 2005.
- A. Griffin and J. R. Hauser, “The voice of the customer,” Marketing Science, vol. 12, no. 1, pp. 1–27, 1993.
- D. X. Cao, Y. H. Han, J. Yang, G. Yang, and C. X. Cui, “Integrated modeling towards collaborative product development,” Advanced Materials Research, vol. 44-46, pp. 669–676, 2008.
- C.-H. Chen, L. P. Khoo, and W. Yan, “Evaluation of multicultural factors from elicited customer requirements for new product development,” Research in Engineering Design, vol. 14, no. 3, pp. 119–130, 2003.
- R. N. Bolton, “A dynamic model of the duration of the customer's relationship with a continuous service provider: the role of satisfaction,” Marketing Science, vol. 17, no. 1, pp. 45–65, 1998.
- R. Decker and M. Trusov, “Estimating aggregate consumer preferences from online product reviews,” International Journal of Research in Marketing, vol. 27, no. 4, pp. 293–307, 2010.
- P. E. Green and V. Srinivasan, “Conjoint analysis in marketing: new developments with implications for research and practice,” Journal of Marketing, vol. 4, pp. 3–19, 1990.
- P. L. Hauge and L. A. Stauffer, “ELK: a method for eliciting knowledge from customers,” in Proceedings of the ASME Design and Methodology, pp. 73–81, September 1993.
- S. Deerwester, S. T. Dumais, G. W. Furnas, T. K. Landauer, and R. Harshman, “Indexing by latent semantic analysis,” Journal of the American Society for Information Science, vol. 41, no. 6, pp. 391–407, 1990.
- M. C. Yang and H. Ji, “A text-based analysis approach to representing the design selection process,” in Proceedings of the International Conference on Engineering Design (ICED ’07), pp. 28–31, Cite Des, Paris, France, August.
- D. Cao, Z. Li, and K. Ramani, “Ontology-based customer preference modeling for concept generation,” Advanced Engineering Informatics, vol. 25, no. 2, pp. 162–176, 2011.
- R. Mugge, P. C. M. Govers, and J. P. L. Schoormans, “The development and testing of a product personality scale,” Design Studies, vol. 30, no. 3, pp. 287–302, 2009.
- A. Chwolka and M. G. Raith, “Group preference aggregation with the AHP: implications for multiple-issue agendas,” European Journal of Operational Research, vol. 132, no. 1, pp. 176–186, 2001.
- P. M. West, P. L. Brockett, and L. L. Golden, “A comparative analysis of neural networks and statistical methods for predicting consumer choice,” Marketing Science, vol. 16, no. 4, pp. 370–391, 1997.
- T.-K. See and K. Lewis, “A decision support formulation for design teams: a study in preference aggregation and handling unequal group members,” in Proceedings of the ASME International Design Engineering Technical Conferences and Computers and Information in Engineering Conference (DETC '05), pp. 177–187, Long Beach, Calif, USA, September 2005.
- R. Studer, V. R. Benjamins, and D. Fensel, “Knowledge engineering: principles and methods,” Data and Knowledge Engineering, vol. 25, no. 1-2, pp. 161–197, 1998.
- Y. Kitamura and R. Mizoguchi, “Ontology-based systematization of functional knowledge,” Journal of Engineering Design, vol. 15, no. 4, pp. 327–351, 2004.
- Z. Li, M. C. Yang, and K. Ramani, “A methodology for engineering ontology acquisition and validation,” Artificial Intelligence for Engineering Design, Analysis and Manufacturing, vol. 23, no. 1, pp. 37–51, 2009.
- M. Marcus, B. Santorini, and M. A. Marcinkiewicz, “Building a large annotated corpus of English: the penn treebank,” Computational Linguistics, vol. 19, no. 2, pp. 313–330, 1994.
- G. A. Miller, “A lexical database for English,” Communications of the ACM, vol. 38, no. 11, pp. 39–41, 1995.
- G. Salton, Automatic Text Process, Addison-Wesley, Wokingham, Mass, USA, 1998.
- J. Z. Wang and W. Taylor, “Concept forest: a new ontology-assisted text document similarity measurement method,” in Proceedings of the IEEE/WIC/ACM International Conference on Web Intelligence (WI '07), pp. 395–401, November 2007.
- C. McMahon, A. Lowe, S. Culley et al., “An integrated search and retrieval system for engineering documents,” Journal of Computing and Information Science in Engineering, vol. 4, no. 4, pp. 329–338, 2004.