A time-sensitive trust information processing method

Abstract

The present invention relates to a time-sensitive trust degree information processing method, including: the interaction information of users interacting with services is stored and calculated in the form of "generations"; the direct trust of service s by each generation of user c in the previous n generations Calculate the direct trust degree of user c to service s by the weighted sum of degrees; calculate the indirect trust degree of user c to service s through the weighted sum of the direct trust degree of user c’s “similar users” on service s in the previous n generations; The weighted sum of c's direct trust degree and indirect trust degree on service s calculates user c's subjective trust degree on service s. Compared with the prior art, the present invention stores and calculates the trust degree through "generation", which effectively reduces the occupied space, introduces the concept of "similar users" and "generation" to calculate the indirect trust degree, and according to the recent generations Dynamically adjust parameters to improve efficiency, reduce time consumption, and reflect time sensitivity.

Description

technical field [0001] The invention relates to the fields of service selection and data mining, in particular to a time-sensitive trust degree information processing method, which can be used to obtain the subjective trust degree of service consumers for services. Background technique [0002] Most of the non-functional requirements currently used for service selection only consider objective factors, but rarely consider subjective factors based on trust or evaluation. In the calculation method of trust, it is necessary to consider the information of historical interaction and make predictions based on it. But saving and calculating the evaluation of each interaction will consume a lot of space and time. At the same time, in order to reduce the impact of malicious evaluation users on the global data, many methods will punish or downgrade users whose evaluation values ​​differ greatly from the overall level, but this also reduces the diversity of preferences and affects the...

AI Technical Summary

Problems solved by technology

But if you save the evaluation of each interaction and calculate it, it will consume a lot of space and time

At the same time, in order to reduce the impact of malicious evaluation users on the global data, many methods will punish or downgrade users whose evaluation values ​​differ greatly from the overall level, but this also reduces the diversity of preferences and affects the prediction of personalized users.

Images