System and method for rule creation from natural language text

A system for rule creation from natural language text, comprising:

• one or more hardware processors;
• and
• one or more memories communicatively coupled to the one or more hardware processors, wherein the one or more hardware processors are configured to execute programmed instructions stored in the one or more memories to:
• grammatically parse a rule statement in natural language text to obtain a parse tree, wherein the parse tree comprises a noun phrase core sub-tree and a verb phrase core sub-tree;
• partition each of the noun phrase core sub-tree and verb phrase core sub-tree into at least one sub-tree;
• extract at least one operator using the at least one sub-tree of at least one of the noun phrase core sub-tree and verb phrase core sub-tree by:
• comparing the verb phrase with a predefined set of operators;
• and
• extracting the at least one operator associated with the verb phrase based on the comparison, wherein the operator is a semantic web rules language (SWRL) operator and a type of operator is one of a commutative operator and a non-commutative operator;
• simultaneously split the rule statement into groups of 3 adjacent words (3-grams) upon removing irrelevant words to match entities of interest with a predefined list of classes and data properties in an ontology representation;
• compare each of the 3-grams with the predefined list of classes and data properties in the ontological representation to extract data properties and class concepts associated with the rule statement, by:
• selecting a member of the predefined list of data properties as a match when more than one word in one of the 3-grams match with the member of the predefined list of data properties;
• and
• selecting a member with least edit distance from the one of the 3-grams as a match when more than one member in the predefined list of data properties match with the one of the 3-grams;
• assign subscripts to the extracted data properties, class concepts and the at least one operator, wherein an order of assigning the subscript to the at least one operator is based on the type of operator;
• and
• create a rule using the data properties, the at least one operator, the class concepts, the associated subscripts and the rule head.

A non-transitory computer readable medium embodying a program executable in computing device for rule creation from natural language text, said program comprising:

• a program code for grammatically parsing a rule statement in natural language text to obtain a parse tree, wherein the parse tree comprises a noun phrase core sub-tree and a verb phrase core sub-tree;
• a program code for partitioning each of the noun phrase core sub-tree and verb phrase core sub-tree into at least one sub-tree;
• a program code for extracting at least one operator using the at least one sub-tree of at least one of the noun phrase core sub-tree and verb phrase core sub-tree by:
• comparing the verb phrase with a predefined set of operators;
• and
• extracting the at least one operator associated with the verb phrase based on the comparison, wherein the operator is a semantic web rules language (SWRL) operator and a type of operator is one of a commutative operator and a non-commutative operator;
• a program code for simultaneously splitting the rule statement into groups of 3 adjacent words (3-grams) upon removing irrelevant words to match entities of interest with a predefined list of classes and data properties in an ontology representation;
• a program code for comparing each of the 3-grams with the predefined list of classes and data properties in the ontological representation to extract data properties and class concepts associated with the rule statement, by:
• selecting a member of the predefined list of data properties as a match when more than one word in one of the 3-grams match with the member of the predefined list of data properties;
• and
• selecting a member with least edit distance from the one of the 3-grams as a match when more than one member in the predefined list of data properties match with the one of the 3-grams;
• a program code for assigning subscripts to the extracted data properties, class concepts and the at least one operator, wherein an order of assigning the subscript to the at least one operator is based on the type of operator;
• a program code for creating a rule head based on the extracted data properties and the at least one operator;
• and
• a program code for creating a rule using the data properties, the at least one operator, the class concepts, the associated subscripts and the rule head.