LONG-TERM TREND PREDICTION METHOD BASED ON NETWORK HOTSPOT SINGLE-PEAK TOPIC PROPAGATION MODEL

摘要

#$%^&*AU2019100854A420190905.pdf#####ABSTRACT The present invention discloses a long-term trend prediction method based on network hotspot single-peak topic propagation model which includes the following steps: 1) detecting in real time characteristics of persons participating in discussion of a target network hotspot single-peak topic over the network, thereby obtaining a quantity R(t) of potential single-peak topic discussers at a time t, a quantity D(t) of single-peak topic discussers at the time t and a quantity E(t) of single-peak topic discussion quitters at the time t; 2) establishing differential equations for the quantity R(t) of potential singlepeak topic discussers, the quantity D(t) of single-peak topic discussers and the quantity E(t) of single-peak topic discussion quitters with respect to the time t, thereby establishing a network single-peak topic propagation model for reflecting a propagation law of the target network hotspot single-peak topic; 3) based on the network singlepeak topic propagation model of the step 2), providing a network single-peak topic long-term trend prediction method. The present invention can intuitively and quickly predict whether the long-term trend of network single-peak topic is to be weakly persistent or dying out.1/1 FIG. 1 circle_2 - circle N j -N 7 ..- DD(t) "N - .' ~ N circle-3 FIG.23