Knowledge Construction And Query Application Based On Tree Classification

Data based on tree classification widely exists in people's production and life.It has inherent logic and science,and can better represent the inclusion relationship in real things.But limited by its own partition rules,it has some limitations.The main manifestation is that when new partitioning requirements arise,some new classes spanning multiple subclasses will be generated under the original tree classification.The data contained in this new class may be distributed in the branches of the original tree classification.Therefore,it is necessary to construct knowledge from tree classification data and extract these branches to meet the new partitioning requirements.At the same time,this paper designs and implements the function of classified query based on the extracted new class information.In the process of knowledge construction based on tree classification data,this paper starts from the structure of tree classification,and takes the location information of data as the basis of knowledge construction.When it is necessary to construct knowledge according to new class names,the node information matching the new class names is pruned,and these node information is matched and pruned again in the data of tree classification.Repeat this part until the non-node matching is successful,the knowledge construction is completed.By extracting sub-nodes several times,we can extract the information that does not contain new class names but is related to their semantics,avoid the incomplete result of string matching based on new class names,and realize the construction of semantic knowledge based on pruning.In the process of classified query,data need to be pre-processed and transformed into vector form that can be processed by classified query algorithm.Then we need to compare the commonly used classification algorithms and select the classification query algorithm suitable for this paperIn order to verify the usability of the system,this paper tests the knowledge construction and classification query module based on tree classification data in the National Library of Medicine of the United States,and presents the test results more intuitively on the web page.The test results show that the knowledge construction and classification query module based on tree classification can meet the new requirements of partition and query,and the performance is good.