The threats to people’s health from chronic diseases are always exist and increasing gradually. How to decrease these threats is an important issue in medical treatment. Thus, this paper suggests a model of a chronic diseases prognosis and diagnosis system integrating data mining (DM) and case-based reasoning (CBR). The main processes of the system include: (1) adopting data mining techniques to discover the implicit meaningful rules from health examination data, (2) using the extracted rules for the specific chronic diseases prognosis, (3) employing CBR to support the chronic diseases diagnosis and treatments, and (4) expanding these processes to work within a system for the convenience of chronic diseases knowledge creating, organizing, refining, and sharing. The experiment data are collected from a professional health examination center, MJ health screening center, and implemented through the system for analysis. The findings are considered as helpful references for doctors and patients in chronic diseases treatments.
Medical treatment is facing a challenge of knowledge discovery from the growing volume of data. Nowadays enormous data are collected continuously through health examination and medical treatment (Tan, Yu, Heng, & Lee, 2003). How to effectively share experts’ knowledge (or experience) is also another important challenge in medical treatments (Cios and Moore, 2002, Gunnlaugsdottir, 2003 and Hendriks and Vriens, 1999). The threats from chronic diseases to people’s health are always exist and increasing gradually. How to decrease these threats is an important issue in medical treatment.
Classification rules are typically usefully for medical problems that have been applied particularly in the area of medical diagnosis (Freitus, 2002). Additionally, numerous machine-learning (ML) techniques have been applied to the field of medical treatments over the past few decades, such as artificial neural networks, genetic algorithm, fuzzy sets, inductive logic programming, and so on (Becerra-Fernandez, 2000, Cios and Moore, 2002, Evans, 1999, Kukar et al., 1996, Sacha and Cios, 2000, Schmidt and Gierl, 2001 and Setiono, 1996). Most of these applications are particular and involve individual ML technique only, such as using data mining (DM) in a medical diagnosis (Alonso, Caraca-Valente, Gonzalez, & Monte, 2002), extracting rules from pruned neural networks for breast cancer diagnosis (Setiono, 1996), and machine learning in prognosis of the femoral neck fracture recovery (Kukar et al., 1996). A weakness of these applications is lack of systematical integration that is a critical factor to enhance the performance of applying ML in medical treatments.
Therefore, this paper proposes a model of a chronic diseases prognosis and diagnosis (CDPD) system integrating DM and case-based reasoning (CBR). The main processes of the system include (1) adopting data mining techniques to discover the implicit meaningful rules from health examination data, (2) using the extracted rules for the specific chronic diseases prognosis, (3) employing CBR to support the chronic diseases diagnosis and treatments, and (4) expanding these processes to work within a system for the convenience of chronic diseases knowledge creating, organizing, refining, and sharing. The experiment data are collected from a professional health examination center, MJ health screening center, and implemented through the system for analysis.
The rest of this paper is organized as follows. Reviews of related work are first described in Section 2. Section 3 describes the CDPD system architecture. We then present the knowledge creating methodology in Section 4. Section 5 briefs on how the knowledge inferring methodology to predict the probabilities of specific chronic diseases. System implementation and verification results are reported in Section 6. Finally, we make our conclusion and discuss future work in Section 7.
In this paper, we systematically integrate DM techniques with CBR in a model of a chronic diseases prognosis and diagnosis (CDPD) system that includes two processing phases: a knowledge creating phase and a knowledge inferring phase. In the knowledge creating phase, data mining techniques, the decision tree induction algorithm and the case association, are adopted to discover the implicit meaningful rules from health examination data. The extracted rules that are stored in a rule base will be used for the specific chronic diseases prognosis. Basing on the discovered rules, the suffering probabilities of chronic diseases in a new case can be prognosticated. Then, the new case will trigger the CBR mechanism to retrieve the most similar case from the case library for supporting the chronic disease treatment. The health examination data are collected through a professional health examination center, MJ health screening center, and implemented through the system for testing the functionalities and feasibility of the system.
This paper makes five critical contributions: (1) it suggests a systematical method of integrating DM techniques with CBR, (2) it shows that helpful implicit rules are discovered from health examination data through DM techniques, (3) it also shows that the discovered rules from health examination data are helpful for chronic disease prognosis, (4) it furnishes the CBR that retrieves the most similar case from the case library for solving the chronic disease problems of the new case, and (5) the results proved that the CDPD system can act as a medical expert system for discovering the useful rules from health examination data for supporting chronic diseases prognosis and diagnosis. A group of experienced chronic diseases relevant doctors who we interviewed also approved this paper’s contributions.
In the future research, more machine learning techniques, such as neural networks, genetic algorithms, and so forth, could also be applied in this field. Certainly, researchers may expand the system to deal with more chronic diseases.
