Research On The Construction Method Of Enterprises' Technological Innovation Index System Based On Text Mining

Technological innovation is the foundation of the survival and development of an enterprise.It is also the driving force and guarantees for the economic development of the entire society and country.It is significant for an enterprise to discover problems,seek countermeasures,and obtain a competitive advantage by correctly analyzing and evaluating the technological innovation capability.In the traditional enterprise technology innovation evaluation system research,methods such as brainstorming,listed group decision-making,and Delphi are used to determine the influencing factors of enterprise technology innovation capabilities based on experts' knowledge and experience and to construct an evaluation index system.This traditional method is often affected by the subjective factors of expert opinions.Different evaluators will construct different index systems for the same enterprise's technological innovation capability and often have very different results based on different background knowledge and experience.The evaluation results become more subjective,and it is difficult to objectively and comprehensively reflect the influencing factors of the enterprise's technological innovation capabilities.According to the research,many enterprises have accumulated various technical documents and data in the development process,containing the technological innovation's critical information.How mining information from these massive unstructured and structured data,extracting clues,forming methods,and providing auxiliary decision-making for enterprise technological innovation is a significant challenge in the era of big data.Therefore,because of the subjectivity,one-sidedness,and insufficient update timeliness of index selection results of the current research,based on the collected text data of more than 400 enterprises' technological innovations,this paper combines cross-scientific ideas and methods based on research of the topic related to big data and enterprises'technological innovation and integrates corporate technological innovation theories,knowledge management theories with text mining,machine learning,ontology theory,semantic network,the research proposes the construction method of a text mining-based index system in enterprise technology innovation.It conducts a comprehensive and objective study of the factors affecting enterprise technology innovation capabilities.In summary,the main research content and results of this article are as follows.(1)A method framework for constructing an enterprise's technology innovation index system is proposed.First of all,this paper performs natural language preprocessing on the collected unstructured text data related to enterprise technology innovation.It uses the semi-automatic domain ontology building module and the semantic representation module based on the domain ontology to represent the text data.On this basis,text mining is used to carry out text clustering,text classification,and text association analysis based on conceptual semantic models to realize the construction of an index system in the field of enterprise technology innovation,including data collection,knowledge representation,and knowledge mining analysis.(2)This paper constructs enterprise technology innovation's domain ontology and represents semantic concepts based on the domain ontology.This paper uses the bibliometrics method to extract keywords and constructs the enterprise technology innovation's seed domain ontology.An automatic extension method for the enterprise technology innovation ontology based on the LDA topic model is proposed,which enriches and improves the domain ontology knowledge.The text keywords are replaced with the concept of the domain ontology,which strengthens the semantic feature description of concepts and improves text representation accuracy.(3)A method for clustering analysis of knowledge in enterprise technology innovation based on the semantic conceptual models is proposed.It uses the constructed enterprise technology innovation domain ontology to map keywords to the domain ontology concepts.The semantic similarity and relevance matrix between concepts are obtained through the proposed semantic similarity and relatedness algorithm based on the domain ontology.The keyword clustering based on the semantic similarity and relatedness matrix of the concepts is performed.The knowledge clustering analysis method in the field of enterprise technology innovation based on the semantic conceptual model proposed in this paper solves the problem of semantic relations between concepts and realizes the clustering of key factors in enterprise technology innovation.The clustering results can be used as the first-level indicators of the enterprise technology innovation index system.(4)A method for classification and analysis of knowledge in enterprise technological innovation based on the semantic conceptual model is proposed.It takes the knowledge clustering result as the target category set and calculates the weighted maximum value of the semantic similarity and relatedness between the text's key feature words and the categories to obtain the text category.The proposed knowledge classification method based on the semantic conceptual model can solve semantic relations between concepts and realize the knowledge classification of the unlabeled text data sets in enterprise technology innovation.The classification results can be used as the second-level and third-level indicators of the enterprise technology innovation index system.(5)A method for association analysis of knowledge in enterprise technological innovation based on the semantic conceptual model is proposed.It uses the constructed enterprise technology innovation domain ontology and the domain ontology-based semantic distance calculation method to sort the mined association rules by semantic interest and combine the improved FP-Growth algorithm to realize the knowledge association analysis.The knowledge association analysis method based on the semantic conceptual model proposed in this paper is superior in performance to the traditional association analysis method and solves the semantic relationship between concepts,and discovers the potential association relationship between the indicators in the enterprise technology innovation index system.The results can reference companies to discover the interaction mode and importance among the influencing factors of technological innovation capabilities.This paper has carried out the applied research on the construction method of enterprise technology innovation index system.This paper collects relevant data on technological innovation from 400 enterprise technology centers in Beijing and conducts an empirical study with 400 enterprises as an example.The enterprise technology innovation data is preprocessed through the text data collection and preprocessing module.Then the domain ontology is constructed for the enterprise technology innovation,and the semantic concept representation is based on the domain ontology through the semantic concept representation module.Finally,the knowledge mining module based on the semantic conceptual model conducts knowledge clustering analysis,knowledge classification analysis,and knowledge association analysis for the knowledge of enterprise technology innovation.It constructs the Beijing enterprise technology innovation index system and analyzes the potential relationship between the indicators.Through comparison and demonstration with the traditional enterprise technology innovation index system based on the expert group decision-making method,the text mining-based enterprise technology innovation index system proposed in this paper is an objective,comprehensive,and timely updated,which can effectively solve the problem of the subjectivity,one-sidedness,and insufficient update timeliness of index selection results by traditional experts'group decision-making method.