Online services rely on machine identifiers to tailor services such as personalized search and advertising to individual users. The assumption made is that each identifier comprises the behavior of a single person. However, shared machine usage is common, and in these cases, the activity of multiple users may be interwoven within one identifier, creating a potentially noisy signal for applications such as search personalization. In this paper, we propose enhancing Web search personalization applying search activity attribution methods to connect the current query with search history of the correct person. Using logs containing both person and machine identifiers, and logs from a popular commercial search engine, we learn models to accurately assign observed search behaviors to the correct individual. This information is then used to augment existing personalization methods that are currently based only on machine identifiers. We show that we can use this new capability to improve the performance of existing personalization methods. Our promising early findings have implications for the design of support to more accurately individualize the search experience.

@inproceedings{singla14enhancing,
	author = {Adish Singla and Ryen W. White and Ahmed Hassan and Eric Horvitz},
	booktitle = {Proc. Special Interest Group On Information Retrieval (SIGIR)},
	month = {July},
	title = {Enhancing Personalization via Search Activity Attribution},
	year = {2014}}