BACKGROUND
Nowadays, the importance of the treatment of neurological diseases in children is well known to everybody. Currently, there are only 60 neurological specialists in Iran. Due to the lack of necessary equipment, we need to send lab. Samples to the countries which have advanced laboratories for diagnosis. Therefore, a diagnostic expert system can significantly prevent deaths of children. In this paper, we design and implement a rule-based expert system in CLIPS language using certainty factors and forward chaining inference in order to help physicians. Naïve Bayes and Logistic algorithm and the experiences of the expert person have been considered to estimate the certainty factors. The accuracy of this system is 94% for detecting diseases. Performances of these two systems are compared and the results confirm that the certainty factors that are specified by the expert physician have 6.2 percent prediction improvement over the one with Naïve Bayes. It is worth noting that an automatic rule generating program has been developed to reduce the time required to produce rules manually and minimize errors.
OBJECTIVE
Nowadays, the importance of the treatment of neurological diseases in children is well known to everybody. Currently, there are only 60 neurological specialists in Iran. Due to the lack of necessary equipment, we need to send lab. Samples to the countries which have advanced laboratories for diagnosis. Therefore, a diagnostic expert system can significantly prevent deaths of children. In this paper, we design and implement a rule-based expert system in CLIPS language using certainty factors and forward chaining inference in order to help physicians. Naïve Bayes and Logistic algorithm and the experiences of the expert person have been considered to estimate the certainty factors. The accuracy of this system is 94% for detecting diseases. Performances of these two systems are compared and the results confirm that the certainty factors that are specified by the expert physician have 6.2 percent prediction improvement over the one with Naïve Bayes. It is worth noting that an automatic rule generating program has been developed to reduce the time required to produce rules manually and minimize errors.
METHODS
Nowadays, the importance of the treatment of neurological diseases in children is well known to everybody. Currently, there are only 60 neurological specialists in Iran. Due to the lack of necessary equipment, we need to send lab. Samples to the countries which have advanced laboratories for diagnosis. Therefore, a diagnostic expert system can significantly prevent deaths of children. In this paper, we design and implement a rule-based expert system in CLIPS language using certainty factors and forward chaining inference in order to help physicians. Naïve Bayes and Logistic algorithm and the experiences of the expert person have been considered to estimate the certainty factors. The accuracy of this system is 94% for detecting diseases. Performances of these two systems are compared and the results confirm that the certainty factors that are specified by the expert physician have 6.2 percent prediction improvement over the one with Naïve Bayes. It is worth noting that an automatic rule generating program has been developed to reduce the time required to produce rules manually and minimize errors.
RESULTS
Nowadays, the importance of the treatment of neurological diseases in children is well known to everybody. Currently, there are only 60 neurological specialists in Iran. Due to the lack of necessary equipment, we need to send lab. Samples to the countries which have advanced laboratories for diagnosis. Therefore, a diagnostic expert system can significantly prevent deaths of children. In this paper, we design and implement a rule-based expert system in CLIPS language using certainty factors and forward chaining inference in order to help physicians. Naïve Bayes and Logistic algorithm and the experiences of the expert person have been considered to estimate the certainty factors. The accuracy of this system is 94% for detecting diseases. Performances of these two systems are compared and the results confirm that the certainty factors that are specified by the expert physician have 6.2 percent prediction improvement over the one with Naïve Bayes. It is worth noting that an automatic rule generating program has been developed to reduce the time required to produce rules manually and minimize errors.
CONCLUSIONS
Nowadays, the importance of the treatment of neurological diseases in children is well known to everybody. Currently, there are only 60 neurological specialists in Iran. Due to the lack of necessary equipment, we need to send lab. Samples to the countries which have advanced laboratories for diagnosis. Therefore, a diagnostic expert system can significantly prevent deaths of children. In this paper, we design and implement a rule-based expert system in CLIPS language using certainty factors and forward chaining inference in order to help physicians. Naïve Bayes and Logistic algorithm and the experiences of the expert person have been considered to estimate the certainty factors. The accuracy of this system is 94% for detecting diseases. Performances of these two systems are compared and the results confirm that the certainty factors that are specified by the expert physician have 6.2 percent prediction improvement over the one with Naïve Bayes. It is worth noting that an automatic rule generating program has been developed to reduce the time required to produce rules manually and minimize errors.