I need a research report on the topic hybrid intelligent system for medical applicationill provide research paperRequirements and marking criteria
template.doc I.J. Information Engineering and Electronic Business, 2017, 4, 38-46 Published Online July 2017 in MECS (http://www.mecs-press.org/) DOI: 10.5815/ijieeb.2017.04.06 Copyright © 2017 MECS I.J. Information Engineering and Electronic Business, 2017, 4, 38-46 Data Mining Based Hybrid Intelligent System for Medical Application Adane Nega Faculty of Computing, Bahir Dar University, Ethiopia Email: nega2002@gmail.com Alemu Kumlachew Faculty of Computing, Bahir Dar University, Ethiopia Email: alemupilatose@gmail.com Abstract—Hybrid intelligent system is a combination of artificial intelligence (AI) techniques that can be applied in healthcare to solve complex medical problems. Case- based reasoning (CBR) and rule based reasoning (RBR) are the two more popular AI techniques which can be easily combined. Both techniques deal with medical data and domain knowledge in diagnosing patient conditions. This paper proposes a hybrid intelligent system that uses data mining technique as a tool for knowledge acquisition process. Data Mining solves the knowledge acquisition problem of rule based reasoning by supplying extracted knowledge to rule based reasoning system. We use WEKA for model construction and evaluation, Java NetBeans for integrating data mining results with rule based reasoning and Prolog for knowledge representation. To select the best model for disease diagnosis, four experiments were carried out using J48, BFTree, JRIP and PART. The PART classification algorithm is selected as best classification algorithm and the rules generated from the PART classifier are used for the development of knowledge base of hybrid intelligent system. In this study, the proposed system measured an accuracy of 87.5% and usability of 89.2%. Index Terms—Hybrid intelligent system, Data mining, Rule based reasoning; Case based reasoning, knowledge acquisition. I. INTRODUCTION Computer systems have an increasing impact on the practice of medicine. Artificial intelligence (AI) system is one of the promising and ambitious areas of information technology to improve human health and longevity. AI has a number of important medical applications, such as modeling brain functions, speech analysis and synthesis, patient monitoring, medical diagnostic systems, drug dosage administration and health care services [19]. Apart from medical applications, Ioannis and JIM [6] described that AI can be used in the design of intelligent tutoring system, which uses an expert system to make decisions during the teaching process. In medical diagnosis, AI is a study realized to emulate human intelligence into computer technology that could assist both doctors and patients by providing a laboratory for examination, representation and cataloguing of medical information as well as by devising novel tools to support decision making and research. The increased integration of intelligent AI techniques in everyday medical applications could improve the efficiency of diagnosis and treatments services by supporting patients and healthcare practitioners [20].There are various ways of solving problems using artificial intelligence, the two more popular techniques are case-based reasoning (CBR) and rule based reasoning (RBR). Both techniques deal with medical data and domain knowledge in diagnosing patient conditions as well as recommending suitable treatments for the particular patients. They serve to improve the quality of medical decision-making, increases patient compliance, minimize complications and medical errors [21].Rule based reasoning and case based reasoning systems can be easily combined to form a hybrid intelligent system. In hybrid intelligent systems, each AI technique has its own strengths and weaknesses.RBR depends on basic rules and regulations of the relevant field and the knowledge from the experts. However, CBR is a way of solving new problems that can adapt to conditions without needing the help of experts [22]. The cost and performance of intelligent system depends directly on the quality of acquired knowledge. The traditional approach to knowledge acquisition is time consuming, costly and error prone as it involves a minimum of two expensive people to communicate i.e. the domain expert and the knowledge engineer. Knowledge acquisition is one of the greatest bottlenecks in the development of hybrid intelligent systems. This is due to that the human expert will usually have insufficient knowledge about intelligent system techniques and the expert will find it difficult to describe his knowledge completely and correctly [24].In order to solve the traditional knowledge acquisition problems and to enrich the knowledge base, data mining techniques, more general, knowledge discovery techniques can be integrated with the hybrid intelligent systems. Data mining (DM) is a subfield of machine learning that enables finding interesting knowledge (patterns, models and relationships) in very large databases [25]. mailto:alemupilatose@gmail.com Data Mining Based Hybrid Intelligent System for Medical Application 39 Copyright © 2017 MECS I.J. Information Engineering and Electronic Business, 2017, 4, 38-46 This paper presents a hybrid intelligent system that uses data mining technique as a tool for knowledge acquisition. The study focuses on applying data mining algorithms to a medical database of Tuberculosis (TB) and uses the database resulting of the mining process in a hybrid intelligent system that will help in medical diagnosis and treatment. The proposed system continues our previous work [24] by considering an automatic knowledge acquisition approach (i.e. using data mining) in the development of hybrid intelligent system. Data mining is the extraction of interesting and previously unknown information or patterns from data sources [26]. It can applicable in diverse areas such as biological data analysis [36], financial data analysis [37], and weather forecasting [38] and so on. Data mining is the central point of knowledge discovery in databases (KDD) process and it correspond to the modeling step in the knowledge discovery in databases process. It involves the application of intelligent methods in order to discover new and useful patterns from large volumes of data. Several models for KDD process have been proposed, but the most known is the industrial model - CRISP-DM. Accordingly to this model, KDD is an iterative and interactive process consisting of six steps: business understanding, data understanding, data preparation, modeling, evaluation of the model and deployment [23]. Data mining applications may solve two kinds of problems: prediction and knowledge discovery [11] [25].For each of these problems it is indicated to use some associated methods. For prediction, we may use classification or regression, while for knowledge discovery we may use clustering, association rules, database segmentation, sequence analysis or visualization. A classification rule attempts to predict the value of a discrete dependent variable from various known attributes. One of the most frequently used methods is classification based on decision tree. The decision tree can predict a new data instance, by following a path that starts from the root to a leaf node. One of the advantages of decision trees lies in the fact that they can easily translate into a set of ‘IF –THEN’ rules, easier to understand. Clustering, often referred to as unsupervised learning, involve a process that discovers structures in data without any supervision. As the name clustering implies, unsupervised algorithm is able to discover structures on its own, by exploiting similarities or differences between individual data points on a data set. Association rules mining is an important data mining method that aims to find interesting dependencies in large sets of data items. Interesting associations between data items can lead to information used for decision making. There are three basic categories of approaches for integrating rule-based reasoning (RBR) with case-based reasoning (CBR). The categorization is based on the importance of each of the two component schemes in the inference process [5]. Rule-Dominant Approach: This approach focuses on the rule-based component and invokes the case-based component only when rules are unable to deal with specialized situations. This augmentation is done by taking the rules as a starting point of problem-solving and then invoking case-based reasoning to handle exceptions to the rules. Case-dominant approach: Here, the CBR module comes first followed by the RBR module. In this paradigm, the rules play a supportive role to case-based reasoning, useful for instance when the case library contains a limited number of cases. Balanced approach: Balanced approaches follow a ‘mixed’ paradigm, where the invocation order of the integrated components is not preset and usually during inference one component dynamically calls the other and vice versa. For this study, the authors use rule-dominant approach. The main reasons for adopting rule-dominant approach as demonstrated by [2] are acceptable accuracy of the inference process, good explanatory ability and the convenient knowledge acquisition process. Moreover, this approach allows the cases and rules to be stored separately in the knowledge base which makes the system a lot easier to be maintained and modified whenever it is needed. II. RELATED WORKS There are a many researches that have been done in the area of data mining, knowledge based systems and hybrid intelligence systems for improving healthcare services. Sellppan and Rafiah [30] have developed a prototype intelligent heart disease prediction system using data mining techniques, namely, decision tree, naïve bayes and neural network. The system uses medical profile such as age, sex, blood pressure and blood sugar to predict the likelihood of patients getting a heart disease. Kapil and Durga [12] presented a general hybrid framework that can support design, planning and management of long term medical conditions. They propose a combination of model-based, case based and rule-based reasoning. The model based reasoning is used as separate reasoning method only if the combination of case-based and rule- based methods is unable to suggest a solution. I.G.L. da Silva et al. [27] proposed an integration of data mining and hybrid expert system. The main goals of this work was to apply knowledge discovery in databases (KDD)