Microsoft Technology Licensing, LLC

Prateek Janardhan

Krishnaram Kenthapadi

Ryan Wade Sandler

Lu Zheng

Kedar U Kulkarni

Mads Johnsen

The disclosed embodiments provide a system for processing data. During operation, the system determines features related to an application for an opportunity by a member of an online network, wherein the features include a historical response rate for a poster of the opportunity and a submission number of the application for the opportunity. Next, the system applies a machine learning model to the features to produce a score representing a likelihood of the member receiving a response to the application from the poster. The system then compares the score to a threshold to determine a recommendation related to applying to the opportunity by the member. Finally, the system outputs the recommendation to improve an experience of applying to the opportunity.

Interaction-based predictions and recommendations for applicants

In an embodiment, a cohort to target is identified, the cohort including a plurality of members of a social networking service having member profiles that all share at least one attribute value. A minimum number of eligible members needed to provide relevant statistical insights from confidential data submitted by eligible members of the cohort is identified. Then, based on an assumed response rate for eligible members of the cohort, a total number of eligible members in the cohort, and the minimum number of eligible members needed to provide relevant statistical insights, it is determined that an estimated amount of responses to invitations to submit confidential data to eligible members of the cohort is less than the minimum number of eligible members needed to provide relevant statistical insights. In response to the determination, the cohort is altered to include eligible members from at least one other cohort, and the process repeated.

Strategy for collection flow and seeding across cohorts

In an example embodiment, a cohort to target is identified, the cohort including a plurality of members of a social networking service having member profiles that all share at least one attribute value. A minimum number of eligible members of the cohort in order to provide relevant statistical insights from confidential data submitted by eligible members of the cohort is identified, and based on an assumed response rate for eligible members of the cohort, a total number of eligible members in the cohort, and the minimum number of eligible members of the cohort in order to provide relevant statistical insights, the eligible members are divided into a first wave and a second wave.

Growing community of inventors

San Francisco, CA, United States of America

Prateek Janardhan