Global optimization of case-based reasoning for breast cytology diagnosis

Department of Management Information Systems, Dongguk University, 3-26 Pil-Dong, Chung-Gu, Seoul 100-715, South Korea
Expert Systems with Applications (Impact Factor: 2.24). 01/2009; 36(1):724-734. DOI: 10.1016/j.eswa.2007.10.023
Source: DBLP


Case-based reasoning (CBR) is one of the most popular prediction techniques in medical domains because it is easy to apply, has no possibility of overfitting, and provides a good explanation for the output. However, it has a critical limitation – its prediction performance is generally lower than other AI techniques like artificial neural networks (ANN). In order to obtain accurate results from CBR, effective retrieval and matching of useful prior cases for the problem is essential, but it is still a controversial issue to design a good matching and retrieval mechanism for CBR systems. In this study, we propose a novel approach to enhance the prediction performance of CBR. Our suggestion is the simultaneous optimization of feature weights, instance selection, and the number of neighbors that combine using genetic algorithms (GA). Our model improves the prediction performance in three ways – (1) measuring similarity between cases more accurately by considering relative importance of each feature, (2) eliminating useless or erroneous reference cases, and (3) combining several similar cases represent significant patterns. To validate the usefulness of our model, this study applied it to a real-world case for evaluating cytological features derived directly from a digital scan of breast fine needle aspirate (FNA) slides. Experimental results showed that the prediction accuracy of conventional CBR may be improved significantly by using our model. We also found that our proposed model outperformed all the other optimized models for CBR using GA.

17 Reads
  • Source
    • "The flow of CBR process in solving problems is defined in four steps of RE, including (1) REtrive, taking issue/most similar cases, (2) REuse, reusing problems/case to try to solve the problem, (3) REvise, revising the proposed solution if necessary, and (4) REtain, saving the new solution as the part of the problem / new cases. Fig. 1 CBR Cyclus by Aamodt dan Plaza [31] Fig. 1 CBR Cyclus by Aamodt dan Plaza [7] One strategy in CBR is SBR that has been widely used in various application domains of CBR, such as medical diagnosis [5] and product recommendations [6] to predict the cases related to the new problem. This is typically performed through k-nearest neighbor retrieval or simply k-NN [2]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: In Case Based Reasoning (CBR), retrieval phase is one of the important phases in view of the dependence of the overall effectiveness of the CBR system on the case retrieval stage. To run the process of finding a new case, CBR systems typically utilize knowledge similarity called Similarity-Based Reasoning (SBR) in which the knowledge encoded in the form of term measures is used to calculate the similarity between a new case with the old one. In this paper, we attempt to build a new concept of similarity text for retrieval case on Indonesian medical sentences. The method to be used is the knowledge base of similarity in which the stage included: (i) the utilization of the knowledge for decision-case association, and (ii) the association extraction of knowledge by creating a rule based on attribute data to generate a subset of cases and solutions in a number of cases. The test results then showed the highest values found in the case of the active sentence form at 88.18% for precision, 88.23% for recall and 89.12% for F1-Measure.
    International conference on control, electronic, renewable energy, and communications (ICCEREC 2015), Bandung, Indonesia; 08/2015
  • Source
    • "Cancer Diagnosis (Glez-Pena, et al. 2009) Cancer Classification (Marling, Shubrook and Schwartz 2008) Diabetes (Ahn and Kim 2009) Breast Cancer Diagnosis (Obot and Uzoka 2009) Hepatitis (Ahmed, et al. 2009) "
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper present the brief overview of different CDSSs proposed in order to facilitate the medical practitioners during diagnosis phase, with this we proposed an online KBCDSS. Using this system the medical practitioners of different medical domains gather over one platform from where they can check and verify the reliability of their decision making during diagnosis phase. Aim of the research is to provide guidance to the new medical practitioners as well as to experienced clinicians. Database management systems used as knowledge representation scheme and Case based reasoning technique is applied as an inference mechanism. This research performed data analysis on the Wisconsin breast cancer data set from UCI Machine Learning Repository and implements this medical data set on KBCDSS tool. Keywords: Clinical Decision Support Systems (CDSS), Knowledge based CDSS (KBCDSS), Knowledge Representation (KR), Database Management Systems (DBMS), Case-based reasoning (CBR), Breast Cancer.
  • Source
    • "In their study, vertical and horizontal dimensions of the research data are reduced through their proposed research model, the hybrid feature and instance selection process using genetic algorithms. Ahn and Kim [6] propose a novel approach to enhance the prediction performance of CBR by putting forward the suggestion of the simultaneous optimization of feature weights, instance selection , and the number of neighbors using genetic algorithms (GA). Sabum et al. [7] construct a radial basis function network (RBFN) composed of representative cases created by K-means clustering. "

Show more