Patent · US Active

Retuning of random classification forests to improve efficacy

US10320834B1 · kind B1 · utility

1Cited by
1References
20Claims
0Family size

Assignee

Inventors

Key dates

Filing dateJun 27, 2016
Grant dateJun 11, 2019
Priority date
Expiry dateFeb 20, 2037

Classification

  • Technology area (CPC H)Electricity
  • CPC primaryH04W12/08
  • WIPO fieldDigital communication
  • WIPO sectorElectrical engineering

Abstract

Methods and apparatus for optimizing computer detection of malware using pattern recognition by refreshing random classification forests are described. In one embodiment, the method may include building a random forest with two or more binary decision trees based at least in part on a first set of categorized data, sending the random forest to a client device with a first random forest control value, identifying a second set of categorized data different from the first set of categorized data, calculating a second random forest control value based on the second set of categorized data and sending the second random forest control value to the client device.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.