Analysis Of Affective Tendency Based On Military Ontology

With the development of the Internet age,the explosive growth of network information,can the relevant organizations determine the emotional orientation of network information quickly and accurately,can it respond to the Internet public opinion in a timely manner,so that the network can play an important role in the development to right direction.The sentiment orientation relationship of network information in the public opinion in the military field is more complicated,and it is important to determine the emotional orientation quickly and accurately.Through the analysis of the research methods of the public opinion tendency in the military field,it is found that,based on the statistical methods,used methods to have the advantages of simple implementation but low accuracy,and the machine learning-based method has the advantages of objective knowledge acquisition and high accuracy.Relying on a large training set,the analysis method is based on sentiment correlation has the advantages of analytical precision but limited by the extraction algorithm and semantic relationship.For the problem of the accuracy of emotional text categorization,this paper studies the emotional tendency analysis in the military field.It mainly studies the TF-IDF technology to obtain emotional words in the military field,and then uses the emotional words to combine commonly used emotional ontology,uses HowNet and NTUSD dictionaries to construct the military field.Emotional ontology;studied the calculation of sentiment orientation degree,and converted the calculation of sentiment orientation into the calculation of meaning similarity between sentiment and reference words.It also improved the semantic similarity algorithm of sentiment words,combined with syntactic rules,degree adverbs and Influence of negative words on sentiment tendency,using emotion correlation method to analyze texts;contrast experiments based on emotional ontology based method in military field and dictionary-based method,based on extremum table method and based on emotional ontology method.The experimental results show that the method based on emotional ontology in the military field used in this paper can improve the accuracy of emotional text classification results.The main work of this paper is as follows:(1)Constructing the ontology of emotional vocabulary in the military field.Using TF-IDF technology to count high frequency words unique to the military field,put forward the classification of emotional words in positive and negative categories.Analyze the influence ofturning words,degree adverbs,and negative words on the production of emotionally inclined words,and construct the sub ontology of three types of words.Then use the visual ontology construction tools Protégé and OWL coding language to construct the military domain ontology.(2)An improved algorithm for similarity of emotional words is proposed.SIPO concept semantic similarity algorithm is adopted.From the four dimensions of semantic coincidence,semantic distance,level depth and node density,semantic similarity calculation is considered.An improved semantic distance calculation method is proposed to set a new impact factor.Used to calculate the relationship between the upper and lower bits.Through comparison experiments,the improved algorithm has better results in text classification accuracy.(3)It proposes an emotional lexical similarity algorithm based on positive and negative reference word assignments.The positive and negative reference words are assigned,and the similarity degree algorithm is used to calculate the similarity degree of the constructed words and the reference words in the military domain ontology to obtain the corresponding emotional tendency.(4)Build an analytical platform for sentiment orientation in the military field.The platform implements text segmentation and preprocessing,and divides the text into several emotional units.By matching rules and words,it obtains corresponding rules formulas and propensity values,calculates the overall tendency of the text,and judges the text's emotional orientation.