Yahoo! Inc.

Ozgur Cetin

Rukmini Iyer

Kannan Achan

Erick Cantu-Paz

Eren Manavoglu

Dongwei Cao

A method for combining multiple probability of click models in an online advertising system into a combined predictive model, the method commencing by receiving a feature set slice (e.g. corresponding to demographics or taxonomies or clusters), and using the sliced data for training multiple slice-wise predictive models. The trained slice-wise predictive models are combined by overlaying a weighted distribution model over the trained slice-wise predictive models. The combined predictive model then is used in predicting the probability of a click given a query-advertisement pair in online advertising. The method can flexibly receive slice specifications, and can overlay any one or more of a variety of distribution models, such as a linear combination or a log-linear combination. Using an appropriate weighted distribution model, the combined predictive model reliably yields predictive estimates of occurrence of click events that are at least as good as the best predictive model in the slice-wise predictive model set.

Using linear and log-linear model combinations for estimating probabilities of events

A machine-learning method for estimating probability of a click event in online advertising systems by computing and comparing an aggregated predictive model (a global model) and one or more data-wise sliced predictive models (local models). The method comprises receiving training data having a plurality of features stored in a feature set and constructing a global predictive model that estimates the probability of a click event for the processed feature set. Then, partitioning the global predictive model into one or more data-wise sliced training sets for training a local model from each of the data-wise slices, and then determining whether a particular local model estimates probability of click event for the feature set better than the global model. A given feature set may be collected from historical data, and may comprise a feature vector for a plurality of query-advertisement pairs and a corresponding indicator that represents a click on the advertisement.

Estimating probabilities of events in sponsored search using adaptive models

Embodiments employ a maximum likelihood estimation (MLE) under a covariance matrix floor constraint to predict missing data from observed data. An MLE solution is obtained for approximately Gaussian distributions under the constraint that the covariance matrix is greater than or equal to a positive-definite matrix. In one embodiment, an offline model estimation is performed using an expectation-maximization (EM) approach to estimate various statistical parameters based on observed data. Then, in an online approach, parameters for various missing CTR data may be predicted based on the offline estimated statistical parameters. A non-limiting, non-exhaustive example using the constrained MLE approach is described for predicting missing click-through rate data useable in selecting an advertisement to display with a search query result.

Maximum likelihood estimation under a covariance constraint for predictive modeling

Sponsored search advertising utilizes a click probability as one factor in selecting and ranking advertisements that are displayed with search results. The probability of click may also be referred to as a predicted click-through rate ('CTR') that may be multiplied by an advertiser's bid for a particular advertisement to rank the display of advertisements. An accurate prediction of the click probability improves the potential revenue that is generated by advertisements in a pay per click system. Other advertising systems may benefit from an accurate and reliable estimate for an advertisement's probability of click in different environments and scenarios.

Growing community of inventors

New York, NY, United States of America

Ozgur Cetin