发明名称 Estimation of closeness of topics based on graph analytics
摘要 Embodiments relate to estimating closeness of topics based on graph analytics. A graph that includes a plurality of nodes and edges is accessed. Each node in the graph represents a topic and each edge represents a known association between two topics. A statistical traversal experiment is performed on the graph. A strength of relations between any two topics represented by nodes in the graph is inferred based on statistics extracted from the statistical traversal experiment.
申请公布号 US9483580(B2) 申请公布日期 2016.11.01
申请号 US201313915524 申请日期 2013.06.11
申请人 INTERNATIONAL BUSINESS MACHINES CORPORATION 发明人 Franceschini Michele M.;Jagmohan Ashish;Lastras-Montano Luis A.;Soares Livio
分类号 G06F17/30;G06N5/04;G06N7/00 主分类号 G06F17/30
代理机构 Cantor Colburn LLP 代理人 Cantor Colburn LLP
主权项 1. A system, comprising: a memory having computer readable computer instructions; and a processor for executing the computer readable instructions to perform a method comprising: accessing a graph comprised of a plurality of nodes and edges, each node representing a topic, and each edge representing a known association between two topics; determining a first probability that, given that an agent has expressed an interest in a single topic represented by a node in the graph, the agent is interested in each of the topics represented by nodes in the graph, the determining comprising: performing a statistical traversal experiment on said graph, the performing including using a generalized form of a matrix eigenvector algorithm that includes a Markov chain specialized to the first topic;inferring a strength of relations between the agent and each of the topics represented by nodes in the graph, the inferring based on statistics extracted from the statistical traversal experiment; andadjusting the inferred strength of relations to account for interests expressed in each of the topics by other agents in a reference population; deriving a second probability that, given that the agent has expressed an interest in two or more topics represented by nodes in the graph, the agent is interested in each of the topics represented by nodes in the graph, calculating an estimate probability by combining the first probability and the second probability using a log-likelihood ratios, wherein lack of interest expressed by the agent in each of the topics is represented as subtraction using the log-likelihood ratios.
地址 Armonk NY US