Standard Case-based Reasoning Research Articles

A data-driven explainable case-based reasoning approach for financial risk detection

The rapid development of artificial intelligence methods contributes to their wide applications for forecasting various financial risks in recent years. This study introduces a novel explainable case-based reasoning (CBR) approach without a requirement of rich expertise in financial risk. Compared with other black-box algorithms, the explainable CBR system allows a natural economic interpretation of results. Indeed, the empirical results emphasize the interpretability of the CBR system in predicting financial risk, which is essential for both financial companies and their customers. In addition, our results show that the proposed automatic design CBR system has a good prediction performance compared to other artificial intelligence methods, overcoming the main drawback of a standard CBR system of highly depending on prior domain knowledge about the corresponding field.

A Data-driven Explainable Case-based Reasoning Approach for Financial Risk Detection

The rapid development of artificial intelligence methods contributes to their wide applications for forecasting various financial risks in recent years. This study introduces a novel explainable case-based reasoning (CBR) approach without a requirement of rich expertise in financial risk. Compared with other black-box algorithms, the explainable CBR system allows a natural economic interpretation of results. Indeed, the empirical results emphasize the interpretability of the CBR system in predicting financial risk, which is essential for both financial companies and their customers. In addition, our results show that the proposed automatic design CBR system has a good prediction performance compared to other artificial intelligence methods, overcoming the main drawback of a standard CBR system of highly depending on prior domain knowledge about the corresponding field.

A Decision Support System for the Prediction of Wastewater Pumping Station Failures Based on CBR Continuous Learning Model

Nowadays the communities are facing the problem of waste and wastewater. While wastewater systems have become more complex, the need for development of sustainable solution for wastewater management emerged. Therefore, the development of a Decision Support System (DSS) for wastewater disposal management became necessary. This paper presents a new DSS for predicting the failure of wastewater pumping stations, the system architecture and its implementation. The prediction model is based on the Case Based Reasoning (CBR) classification method. The standard CBR classification technique has been upgraded with an algorithm for continuous learning. The paper describes the system structure, its connection to the wastewater system, the internal processes involved in the prediction process and the implemented algorithm for continuous learning. Furthermore, the features used in the prediction are indicated along with the achieved results and the method of results evaluation. The test and obtained results indicate that the proposed DSS is efficient and capable of providing very good results in the prediction process.

A case-based reasoning system based on weighted heterogeneous value distance metric for breast cancer diagnosis.

We present the implementation and application of a case-based reasoning (CBR) system for breast cancer related diagnoses. By retrieving similar cases in a breast cancer decision support system, oncologists can obtain powerful information or knowledge, complementing their own experiential knowledge, in their medical decision making. We observed two problems in applying standard CBR to this context: the abundance of different types of attributes and the difficulty in eliciting appropriate attribute weights from human experts. We therefore used a distance measure named weighted heterogeneous value distance metric, which can better deal with both continuous and discrete attributes simultaneously than the standard Euclidean distance, and a genetic algorithm for learning the attribute weights involved in this distance measure automatically. We evaluated our CBR system in two case studies, related to benign/malignant tumor prediction and secondary cancer prediction, respectively. Weighted heterogeneous value distance metric with genetic algorithm for weight learning outperformed several alternative attribute matching methods and several classification methods by at least 3.4%, reaching 0.938, 0.883, 0.933, and 0.984 in the first case study, and 0.927, 0.842, 0.939, and 0.989 in the second case study, in terms of accuracy, sensitivity×specificity, F measure, and area under the receiver operating characteristic curve, respectively. The evaluation result indicates the potential of CBR in the breast cancer diagnosis domain.

A hybrid CBR classification model by integrating ANN into CBR

Case-based reasoning (CBR) is an artificial intelligent approach to problem solving and learning, which understands and extracts knowledge from past cases. However, CBR faces the challenge of assigning weights to the features to measure similarity between cases effectively and correctly. Integration of inherent learning capability of artificial neural networks (ANNs) to help the CBR in attributing the correct and appropriate weights to the features is likely to improve the performance of the standard CBR approach. This paper integrates back propagation neural network (BPNN) into CBR in an innovative way to develop an efficient model for classification tasks. The implementation of integration of NN and CBR for classification tasks is done by building training and testing datasets and optimising NN architecture in terms of number of neurons in hidden layer. This paper investigates the integration of multi-layer BP neural network and CBR. The experimental results obtained with the proposed hybrid model are compared with that of standard CBR, CBR with value difference matrix (VDM) and one existing CBR with BPNN approaches. The superiority of the proposed hybrid CBR model is established to others. The performance of the proposed model is validated with four datasets.

Integrating spatial relations into case-based reasoning to solve geographic problems

Case-based reasoning (CBR) method has been widely used to study environmental and spatial problems since the 1990s. Spatial relations among geographic cases and between case and environment (hereinafter to be referred as spatial relations) were not well considered in most of the previous studies. However, these relations are extremely important in geographic problems solving as spatially closer geographic phenomena are more likely to be similar than those are disperse in space. This paper presents a generic application paradigm based on CBR to solve geographic problems. To better consider the spatial relations, a new component of “Geographic Environment” was added into the standard CBR case representation model. A rough set-based algorithm was used to prune essential spatial relations, which were then used to extract key decision rules and retrieve similar past cases for the new problem. Standard CBR directly accepts the solution that was derived from the past similar cases to solve the new problem. In this study, however, solution was not accepted unless it also satisfied the key decision rules. An illustrating example was used to demonstrate how the general framework and algorithm could be applied to solve geographic problems. The algorithm was then evaluated by examining two datasets, the 2003 land use in the Pearl River Estuary and land use change from 1995 to 2000 in Zhuhai city of China. These two datasets were also examined by the standard CBR and Bayesian Networks (BNs) methods. Comparison of the stratified 10-fold cross-validation results indicated that the algorithm proposed in this study yielded statistically significant higher overall validation accuracy than the standard CBR and BNs methods.

Efficient Case-Based Structure Generation for Design Support

This paper describes a general approach to support case-based structure generation named Conceptual Analogy (Borner 1997). The approach can be used to support design tasks in domains that do not allow the acquisition of a complete and consistent set of constraints or rules but that do provide a larger set of past experiences. The experiences – also called cases – may resemble for example CAD data of room layouts or pipe systems. Thus, cases represent solutions to particular problems without an explicit, a priori separation of problem and solution variables. The paper starts with a description and formalization of the case-based structure generation task and a discussion of why standard case-based reasoning and other approaches are not applicable. Then, the new approach of Conceptual Analogy (CA) is introduced. CA applies conceptual clustering to organize cases represented by graphs into hierarchical classes of structurally similar cases. These case classes are then represented by concepts. Given a problem, i.e., a partial solution, the hierarchy of concepts is searched for applicable concepts, i.e., concepts that allow the generation of at least one solution to the given problem. Applicable concepts are used to generate a set of solutions that can be ordered according to their quality. Properties of the approach as well as complexity results are presented. An architectural design domain and task where the approach has been applied successfully, is used for illustration and for practical evaluation. Finally, the approach and its implementation are compared to two systems that aim at the support of similar design tasks. The paper concludes with an assessment of the future direction of this research.