Determining risk of malware infection in enterprise hosts

1. A method comprising steps of:

obtaining data characterizing host devices of a computer network of an enterprise;

applying the data to a logistic regression model to generate malware infection risk scores for respective ones of the host devices; and

taking one or more proactive measures to prevent malware infection in one or more of the host devices based at least in part on the malware infection risk scores;

wherein the malware infection risk scores indicate likelihoods that the respective host devices will become infected with malware;

wherein the logistic regression model incorporates a plurality of features of the host devices as respective model variables including at least user demographic features, virtual private network (VPN) activity features and web activity features of the host devices;

wherein the user demographic features incorporated by the logistic regression model comprise at least one of user level in enterprise hierarchy and user technical level;

wherein the VPN activity features incorporated by the logistic regression model comprise number of VPN connections and duration of the VPN connections;

wherein the web activity features incorporated by the logistic regression model comprise web usage features and domain reputation features, the web usage features comprising number of distinct domains visited and the domain reputation features comprising number of connections challenged by web proxy and number of connections consented by web proxy;

wherein the data characterizing the host devices comprises data for the plurality of features incorporated by the logistic regression model; and

wherein the steps are performed by at least one processing device comprising a processor coupled to a memory.

2. The method of claim 1 wherein taking one or more proactive measures further comprises:

identifying a subset of the host devices having malware infection risk scores above a specified threshold; and

applying the one or more proactive measures to the host devices in the subset but not to the host devices outside of the subset.

3. The method of claim 1 wherein the user demographic features incorporated by the logistic regression model further comprise one or more of user gender and user geographic location.

4. The method of claim 1 wherein the data characterizing the host devices for the user demographic features is obtained at least in part from one or more employee databases of the enterprise.

5. The method of claim 1 wherein the VPN activity features incorporated by the logistic regression model further comprise one or more of amount of data sent over VPN connections and number of external network addresses of VPN connections.

6. The method of claim 1 wherein the data characterizing the host devices for the VPN activity features is obtained at least in part from VPN logs of the enterprise.

7. The method of claim 1 wherein the web activity features incorporated by the logistic regression model further comprise visited web site category features.

8. The method of claim 7 wherein the visited web site category features comprise one or more of number of chat sites visited, number of file transfer sites visited, number of freeware sites visited, number of gaming sites visited, number of social-networking sites visited, number of streaming sites visited and number of non-categorized sites visited.

9. The method of claim 7 wherein the domain reputation features further comprise one or more of number of connections blocked by web proxy and number of new domains visited.

10. The method of claim 1 wherein the data characterizing the host devices for the web activity features is obtained at least in part from web proxy logs of the enterprise.

11. The method of claim 1 further comprising generating the logistic regression model.

12. The method of claim 11 wherein generating the logistic regression model comprises:

identifying a plurality of potential malware infection related features of the host devices;

obtaining data indicative of actual malware infection of particular ones of the host devices;

determining correlations of the data indicative of actual malware infection with respective ones of the potential malware infection related features;

selecting a subset of the potential malware infection related features based on said correlations; and

configuring the logistic regression model to incorporate the selected subset of the potential malware infection related features as said plurality of features of the host devices incorporated by the logistic regression model.

13. The method of claim 12 wherein obtaining data indicative of actual malware infection comprises obtaining that data at least in part from anti-virus logs of the enterprise.

14. An article of manufacture comprising a processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code when executed by at least one processing device causes said at least one processing device:

to obtain data characterizing host devices of a computer network of an enterprise;

to apply the data to a logistic regression model to generate malware infection risk scores for respective ones of the host devices; and

to take one or more proactive measures to prevent malware infection in one or more of the host devices based at least in part on the malware infection risk scores;

wherein the malware infection risk scores indicate likelihoods that the respective host devices will become infected with malware;

wherein the logistic regression model incorporates a plurality of features of the host devices as respective model variables including at least user demographic features, virtual private network (VPN) activity features and web activity features of the host devices;

wherein the user demographic features incorporated by the logistic regression model comprise at least one of user level in enterprise hierarchy and user technical level;

wherein the VPN activity features incorporated by the logistic regression model comprise number of VPN connections and duration of the VPN connections;

wherein the web activity features incorporated by the logistic regression model comprise web usage features and domain reputation features, the web usage features comprising number of distinct domains visited and the domain reputation features comprising number of connections challenged by web proxy and number of connections consented by web proxy; and

wherein the data characterizing the host devices comprises data for the plurality of features incorporated by the logistic regression model.

15. An apparatus comprising:

at least one processing device comprising a processor coupled to a memory;

said at least one processing device being configured:

to obtain data characterizing host devices of a computer network of an enterprise;

to apply the data to a logistic regression model to generate malware infection risk scores for respective ones of the host devices; and

to take one or more proactive measures to prevent malware infection in one or more of the host devices based at least in part on the malware infection risk scores;

wherein the malware infection risk scores indicate likelihoods that the respective host devices will become infected with malware;

wherein the logistic regression model incorporates a plurality of features of the host devices as respective model variables including at least user demographic features, virtual private network (VPN) activity features and web activity features of the host devices;

wherein the user demographic features incorporated by the logistic regression model comprise at least one of user level in enterprise hierarchy and user technical level;

wherein the VPN activity features incorporated by the logistic regression model comprise number of VPN connections and duration of the VPN connections;

wherein the web activity features incorporated by the logistic regression model comprise web usage features and domain reputation features, the web usage features comprising number of distinct domains visited and the domain reputation features comprising number of connections challenged by web proxy and number of connections consented by web proxy; and

wherein the data characterizing the host devices comprises data for the plurality of features incorporated by the logistic regression model.

16. The apparatus of claim 15 wherein the data characterizing the host devices for the user demographic features is obtained at least in part from one or more employee databases of the enterprise.

17. The apparatus of claim 15 wherein the data characterizing the host devices for the VPN activity features is obtained at least in part from VPN logs of the enterprise.

18. The apparatus of claim 15 wherein the data characterizing the host devices for the web activity features is obtained at least in part from web proxy logs of the enterprise.

19. A network security system comprising the apparatus of claim 15 .

20. The article of manufacture of claim 14 wherein the program code when executed by said at least one processing device further causes said at least one processing device to generate the logistic regression model by:

identifying a plurality of potential malware infection related features of the host devices;

obtaining data indicative of actual malware infection of particular ones of the host devices;

determining correlations of the data indicative of actual malware infection with respective ones of the potential malware infection related features;

selecting a subset of the potential malware infection related features based on said correlations; and

configuring the logistic regression model to incorporate the selected subset of the potential malware infection related features as said plurality of features of the host devices incorporated by the logistic regression model.