The invention relates to a network negative information
impact minimization method based on a
topic model. The network negative information
impact minimization method comprises the following steps: 1) adopting a
directed graph to express the propagation of information in a
social network, and independently calculating the probability distribution of the negative information and the probability distribution of historical information on each edge through the
topic model; 2) independently calculating a distance, i.e., KL (Kullback-Leibler)
divergence d(w,i), between the probability distribution of the negative information and the probability distribution of historical information on each edge, wherein d expresses a calculation result of the KL
divergence, w expresses the topic distribution of the historical information, and i expresses the topic distribution of the negative information; 3) calculating tb(w)=b(w) / d(w,i) and to(w)=o(w) / d(w,i), wherein b(w) and o(w) are independently the calculation results of a centrality
algorithm and an out-degree in-degree
algorithm and are sorted from big to small, and the first k nodes are removed to minimize the propagation range of the negative information. The method disclosed by the invention can effectively control the
social network of which the malicious information is burst, and greatly reduces an
impact range of the negative information.