The paper uses information gain theory to analyze the correlation between each influencing factor and the level of the information cocoons, builds the model of the sensitive influence factor on the level of the information cocoons, and then builds the prediction model of the level of the information cocoons through the support vector machine theory. The 12 sensitive influencing factors of the level of information cocoons are concentrated in four sub-dimensions, including algorithm recommendation technology, user information literacy, and system interactivity and user behavior characteristics. The accuracy of model classification prediction is 84.92%, indicating that the support vector machine has good predictive ability for the level of information cocoons .
Wang Yicheng
,
Wang Ping
,
Wang Meiyue
. Research on SVM-based Network Information Cocoons Hierarchy Sensitive Influencing Factors Identification[J]. Information and Documentation Services, 2019
, 40(6)
: 90
-97
.
DOI: 10.12154/j.qbzlgz.2019.06.012
