Research on Intelligent Question and Answering Based on a Pet Knowledge Map

question: What are the symptoms of nm getting nd? In the above example, the pet’s proper nouns, such as golden retriever, which are involved in the natural language question of the user, will be converted into the golden retriever’s part of speech nm after the entity similarity calculation, and the shift is transformed into the distemper word nf instead. The advantage of this is that it can reduce the selection workload of the naive Bayesian classifier feature. Additionally, because there is no special dataset in the pet field, the workload of building the dataset can be reduced, and the required training set can be reduced in size. The specific conversion is shown in Table 8. 3.8. Text Classification Based on Multiple Naive Bayes This article requires multiple classifications of pet text datasets. At present, there are many machine learning and deep learning Y. Liu et al. / International Journal of Networked and Distributed Computing 8(3) 162–170 169 Table 8 | Rule conversion table Conversion rule User problem Abstract problem Pet breed name — nm Golden retriever price Price of nm Pet disease name — nd What are the symptoms of golden retrievers? What is the symptom of nm? Pet food — nf Can golden retriever eat grapes? Can nm eat nf? algorithms that can perform multi-classification of texts. Multiple naive Bayes have stable classification efficiency and good performance for small-scale data and multi-classification. Because there are very few corpora in the pet field, the size of the corpus built in this paper is also very small, so this paper adopts a naive Bayesian text classifier based on polynomials. Based on the knowledge of pet knowledge maps, a total of 24 categories are constructed according to the pet breed, pet disease and pet food attributes. The user’s natural language question will match one of the 24 categories after multidirectional naive Bayes classification as the classification results. 3.9. Matching Word Sequence Diagram Through the classification result of the text quantifier based on multiple naive Bayes, the labels of the categories corresponding to the natural language problem of the user, such as weight, price, and main symptoms, are obtained, which are labels corresponding to the user problem and correspond to natural language questions. Then, the determined intention tag maps the corresponding question template, matching the word order graph in the template. The natural language question basically describes the relationship between the subject and the object, while the graph model can describe the relationship between the node and the node through the edge. The word map is a directed graph, the subject points to the object, and the predicate is used as an edge. In directed graphs, subjects and objects are entities, and predicates are relationships between entities, including attribute relationships. For example, what are the symptoms of a golden retriever with canine distemper? The conversion into a word sequence diagram is shown in Figure 6. This paper constructs a template for a total of 24 types of problems in three major categories. The problem template example is shown in Table 9. 3.10. Generate an Answer The pet knowledge map is stored in the graph database OrientDB. In this paper, the word map is converted into OrientDB’s SQL query statement, the answer is stored in the graph database OrientDB storing the knowledge map, and the answer to the question is returned to the user. The automatic question answering system based on the pet knowledge map supports the origin, price, IQ, disease overview, symptoms, prevention and other issues of pets and can answer three Figure 6 | Example of a word map. Table 9 | Example of a problem template Question type Problem template Price Nm price The main symptoms Nm has disease nd main symptoms Edible Nm edible nd edible Figure 7 | Price of a golden retriever. major questions in total. As shown in Figure 7, the answer is the question of the pet breed attribute, such as the price of a golden retriever.