Intelligent Fril/SQL interrogator

Abstract

The intelligent Fril/SQL interrogator is an object-oriented and knowledge-based support query system, which is implemented by the set of logic objects linking one another. These logic objects integrate SQL query, support logic programming language-Fril and Fril query together by processing them in sequence in slots of each logic object. This approach therefore takes advantage of both object-oriented system and a logic programming-based system. Fuzzy logic data mining and a machine learning tool kit built in the intelligent interrogator can automatically provide a knowledge base or rules to assist a human to analyze huge data sets or create intelligent controllers. Alternatively, users can write or edit the knowledge base or rules according to their requirements, so that the intelligent interrogator is also a support logic programming environment where users can write and run various Fril programs through these logic objects.

Fuzzy prototype model and semantic distance

“…Intelligent Fril/SQL interrogator [11] is an objectoriented and knowledge-based support query system, which not only integrates SQL with support logic programming language Fril, but also is embedded with fuzzy logic machine learning toolkit, fuzzy partition graphic interface, natural language interface, data visualization tool, knowledge base customization editor, etc. The system flow graph is illustrated in Fig.…”

“…The accuracy of test set is calculated by Eq. (11), and its result is listed in the column of ''Test set'' in Table 1. To guarantee the model keeping the right information, we also calculate the accuracy of the training set using the same method, whose result is also listed in the column of ''Training set'' in Table 1.…”

“…A logic object [11] defined in Fig. 9 is a collection of serial processes to retrieve and output the relations and knowledge required by the user, according to the inputs of this object.…”

“…For fuzzy decision tree proposed in Section 2.3.2, we use this fuzzy sets model for continuous (numerical) attributes, and still use crisp sets as a special case for discrete (nominal) attributes [9][10][11].…”

“…Therefore, we design a hybrid fuzzy expert system to list intelligent product suggestions for these new customers, who have no previous shopping history records in the supermarket application. The system includes both the fuzzy prototype model and semantic distance, and is implemented by intelligent Fril/SQL interrogator [11], which is an object-oriented and knowledge-based support query system introduced in Section 4.4.…”