ARTICLE N PRESS xpert Systems with Applications xxx(2011)xXX-XXX Contents lists available at SciVerse Science Direct Expert Systems with Applications ELSEVIER journalhomepagewww.elsevier.com/locate/eswa A recommender mechanism based on case-based reasoning Chen-Shu Wang Heng-Li Yang Graduate Institute of information and Logistics Management, National Taipei University of Technology. 1, Sec. 3, Chung-hsiao E Road, Taipei, Taiwan Department of mis, National Chengchi University, 64, Sec. 2, Chihnan Road, Taipei, Taiwan ARTICLE INFO ABSTRACT ase-based reasoning(CBr)algorithm is particularly suitable for solving ill-defined and unstructured decision-making problems in many different areas. The traditional CBR algorithm, however, is inappro priate to deal with complicated problems and therefore needs to be further revised. This study thus pr ultiple stage poses a next-generation CBR(GCBR) model and algorithm. GCBR presents as a new problem-solving Artificial intelligence application decision-making problems by using hierarchical criteria architecture(HCA)problem representation which involves multiple decision objectives on each level of hierarchical, multiple-level decision criteria, thereby enables decision makers to identify problems more precisely. Additionally, the proposed GCBR can also provide decision makers with series of cases in support of these multiple decision-making stages. GCBR furthermore employs a genetic algorithm in its implementation in order to reduce the effort involved in case evaluation. This study found experimentally that using GCBR for making travel-planning recommendations involved approximately 80% effort than traditional CBR, and therefore concluded that GCBR should be the next generation of case-based reasoning algorithms and can be applied to actual case-based recommender mechanism implementation. e 2011 Elsevier Ltd. All rights reserved. 1 Introduction Furthermore, Chang(2005)applied CBr to screening children with delayed development in order to detect their disorder early Case based reasoning( CBR)is a paradigm, concept, and intuitive rough analysis of their symptoms mechanism for solving ill-defined and unstructured problems chances of effective treatment. Both Garrell, Golobardes, Bernado ( Belecheanu, Pawar, Barson, Bredehorst, Weber, 2003). Similar and Llora(1999)and Golobardes, Llora, Salamo, and Marti(2002) to the natural human problem-solving process, CBR retrieves past have used Cbr to diagnose breast cancer based on mammary nces for reuse in regard to target problems. Since such pro- biopsy data and micro calcifications, respectively. Additionally ess is likely to need to revise previous-case solutions before Shimazu, Shibata, and Nihei(2001)applied conversational case applying them, CBR then retains successful problem-solving expe- based algorithm( CCBr) to developing a mentor guide for user riences for further reuse(Aamodt Plaza, 1994). This, then, is tra- helpdesk implementation and Shimazu(2002)applied CCBr to ditional CBr,'s 4R processes of retrieve, reuse, revise, and retain automatic-clerk mechanisms and electronic website shopping CBR is therefore a classical artificial intelligence algorithm. assistance. Researchers have historically tended to solve these Many have applied CBr within various problem-solving domains problems by using such mathematical models as regressions but (Aamodt Plaza, 1994; Kolodner, 1993; Shiu Pal, 2001; Waston, these mathematical models involve too many assumptions to be 1997). Cirovic and Cekic(2002)applied CBr to construction pro pplied effectively to real-world problem solving, and CBr seems jects during their preliminary design phase by retrieving historical to be a feasible alternative cases from a historical project database, storing useful case(s)in Researchers have until recently extended CBR applications their construction knowledge base, and then applying the most mechanisms for making recommendations based on previous similar previous case(s) to improve the quality of construction cases. Yang and Wang(2009a)applied the CBr algorithm to infor- designs. Belecheanu et al. (2003)referred to past records in order mation-system project management as a recommender mecha- to reduce information uncertainty in regard to such industrial nism by offering project managers preferences from previous ticul uirements as those involved in new product development, par- cases to help project managers construct new project plans. They larly when employing the concurrent engineering approach. also applied similar mechanisms to travel-schedule planning Edu cators, furthermore, can integrate CBr recommender mechanism g author into e-learning systems to provide learners with reference- E-mail addresses: wangcsentutedu w (C-S. Wang). yanhenccuedutw certification paths(2009b). Such real-world problems as these are usually difficult to formulate within strict mathematical 0957-4174/s- see front matter o 2011 Elsevier Ltd. All rights reserved doi:10.1016/eswa2011.09.1 Please cite this article in press as: Wang, C-S,& Yang. H.-L A recommender mechanism based on case-based reasoning Expert Systems with Application 2011da06 /jeswa.201109161A recommender mechanism based on case-based reasoning Chen-Shu Wang a , Heng-Li Yang b,⇑ aGraduate Institute of Information and Logistics Management, National Taipei University of Technology, 1, Sec. 3, Chung-hsiao E. Road, Taipei, Taiwan bDepartment of MIS, National ChengChi University, 64, Sec. 2, Chihnan Road, Taipei, Taiwan article info Keywords: Recommender mechanism Case-based reasoning Multiple stage reasoning Genetic algorithm Artificial intelligence application abstract Case-based reasoning (CBR) algorithm is particularly suitable for solving ill-defined and unstructured decision-making problems in many different areas. The traditional CBR algorithm, however, is inappro￾priate to deal with complicated problems and therefore needs to be further revised. This study thus pro￾poses a next-generation CBR (GCBR) model and algorithm. GCBR presents as a new problem-solving paradigm that is a case-based recommender mechanism for assisting decision making. GCBR can resolve decision-making problems by using hierarchical criteria architecture (HCA) problem representation which involves multiple decision objectives on each level of hierarchical, multiple-level decision criteria, thereby enables decision makers to identify problems more precisely. Additionally, the proposed GCBR can also provide decision makers with series of cases in support of these multiple decision-making stages. GCBR furthermore employs a genetic algorithm in its implementation in order to reduce the effort involved in case evaluation. This study found experimentally that using GCBR for making travel-planning recommendations involved approximately 80% effort than traditional CBR, and therefore concluded that GCBR should be the next generation of case-based reasoning algorithms and can be applied to actual case-based recommender mechanism implementation. 2011 Elsevier Ltd. All rights reserved. 1. Introduction Case based reasoning (CBR) is a paradigm, concept, and intuitive mechanism for solving ill-defined and unstructured problems (Belecheanu, Pawar, Barson, Bredehorst, & Weber, 2003). Similar to the natural human problem-solving process, CBR retrieves past experiences for reuse in regard to target problems. Since such pro￾cess is likely to need to revise previous-case solutions before applying them, CBR then retains successful problem-solving expe￾riences for further reuse (Aamodt & Plaza, 1994). This, then, is tra￾ditional CBR’s 4R processes of retrieve, reuse, revise, and retain. CBR is therefore a classical artificial intelligence algorithm. Many have applied CBR within various problem-solving domains (Aamodt & Plaza, 1994; Kolodner, 1993; Shiu & Pal, 2001; Waston, 1997). Cirovic and Cekic (2002) applied CBR to construction pro￾jects during their preliminary design phase by retrieving historical cases from a historical project database, storing useful case(s) in their construction knowledge base, and then applying the most similar previous case(s) to improve the quality of construction designs. Belecheanu et al. (2003) referred to past records in order to reduce information uncertainty in regard to such industrial requirements as those involved in new product development, par￾ticularly when employing the concurrent engineering approach. Furthermore, Chang (2005) applied CBR to screening children with delayed development in order to detect their disorder early through analysis of their symptoms, thereby improving the chances of effective treatment. Both Garrell, Golobardes, Bernado, and Llora (1999) and Golobardes, Llora, Salamo, and Marti (2002) have used CBR to diagnose breast cancer based on mammary biopsy data and micro calcifications, respectively. Additionally, Shimazu, Shibata, and Nihei (2001) applied conversational case￾based algorithm (CCBR) to developing a mentor guide for user helpdesk implementation and Shimazu (2002) applied CCBR to automatic-clerk mechanisms and electronic website shopping assistance. Researchers have historically tended to solve these problems by using such mathematical models as regressions but these mathematical models involve too many assumptions to be applied effectively to real-world problem solving, and CBR seems to be a feasible alternative. Researchers have until recently extended CBR applications as mechanisms for making recommendations based on previous cases. Yang and Wang (2009a) applied the CBR algorithm to infor￾mation-system project management as a recommender mecha￾nism by offering project managers preferences from previous cases to help project managers construct new project plans. They also applied similar mechanisms to travel-schedule planning. Edu￾cators, furthermore, can integrate CBR recommender mechanism into e-learning systems to provide learners with reference￾certification paths (2009b). Such real-world problems as these are usually difficult to formulate within strict mathematical 0957-4174/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2011.09.161 ⇑ Corresponding author. E-mail addresses: wangcs@ntut.edu.tw (C.-S. Wang), yanh@nccu.edu.tw (H.-L. Yang). Expert Systems with Applications xxx (2011) xxx–xxx Contents lists available at SciVerse ScienceDirect Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa Please cite this article in press as: Wang, C.-S., & Yang, H.-L. A recommender mechanism based on case-based reasoning. Expert Systems with Applications (2011), doi:10.1016/j.eswa.2011.09.161