Research On Knowledge Discovery Based On Multi-level Domain Ontology Representation

With the development of big data,enterprises have a large amount of data resources on the one hand,but on the other hand,they are facing the problem of data explosion and lack of knowledge.As knowledge is increasingly prominent in the importance of enterprise decision-making,how to discover multi-level knowledge from data resources to support decision-making in different scenarios has become one of the important problems that enterprises urgently seek solutions.Traditional knowledge discovery research ignores the auxiliary role of domain knowledge,which leads to many problems such as excessive technical orientation and a large number of redundant mining results.Considering the constraint and guiding function of domain knowledge on knowledge discovery,this paper introduces domain knowledge into the process of knowledge discovery through knowledge representation of domain knowledge,and studies how to mine multi-level rule knowledge from data set based on domain knowledge.Firstly,aiming at the representation of domain knowledge,this paper presents a multi-level domain knowledge ontology model by using ontology representation.The model is composed of concepts,attributes and instances in the domain,and is divided into base level,concept level and instance level from top to bottom and from concrete to abstract.It describes domain knowledge from multi-level and multi-dimensional dimensions,providing a new idea for domain knowledge ontology construction.Focusing on the specific clothing retail domain,the clothing retail knowledge ontology is constructed,and the visualization of the ontology is realized,which verifies the feasibility of the proposed domain knowledge ontology model.Secondly,for the problem of multi-level rule knowledge discovery based on domain knowledge,this paper combines ontology and association rule mining in terms of specific algorithm,makes three optimization adjustments to the FP-growth algorithm,and proposes an ontology-based multi-level association rule mining algorithm to mine multi-level association rules with ontology information.In order to solve the possible redundancy problem in the mining results,the rule filtering methods are used to filter the three redundant rules in the multi-level association rules,which is beneficial to improve the value of rule knowledge.Finally,taking the specific clothing retail field as an example,this paper makes an example analysis based on the real data set of a certain clothing retail enterprise,and the ontology-based multi-level association rule mining algorithm and rule filtering methods are used to extract multi-level rule knowledge from data sets with the domain knowledge represented by the constructed clothing retail knowledge ontology.In this paper,domain knowledge ontology model and ontology-based multi-level association rule mining algorithm are proposed,which are verified by real data sets.The research in this paper takes the clothing retail field as an example,which can also be referenced in other fields.In addition,the research of this paper not only provides new ideas and methods for domain knowledge representation and ontology-based multi-level association rule mining,but also provides support for perfecting the research of knowledge discovery based on domain knowledge to a certain extent.