Recommendation Model Based on Semantic Features and a Knowledge Graph
Algorithm 1
Obtain the feature set of items.
Input: short texts of items
Output: S - source feature entity set of items
1:S=Ø
2: Use DBpedia Spotlight to annotate named entity words in short texts, and calculate their tagging probability values, and get a set of named entity words according to a given threshold. And
3: for to
4: According to the Wikipedia page of wi, we get the candidate entity set for . And .
5: Calculate the probability of each candidate entity in E(wi) being marked as an entity under the current context conditions, and we select the entity with the largest probability to add to S.
