Episerver Inc.

Optimizely North America Inc.

Spencer Eldon Pingry

Jonathan Bartholomew Mulieri

Laura Kreisberg

Brian Taylor

Various implementations of the invention for predicting customer behavior are described. Various implementations of the invention comprise an embedding component configured to receive and embed sequential inputs regarding a plurality of customer interactions with an online presence of a client; a plurality of causal dilated convolutional 'CDC' elements configured to receive the embedded sequential inputs and to output a feature vector, where each CDC element comprises two causal dilated convolutions with regularization that is bypassed with a skip connection; a plurality of dense neural network elements configured to receive the feature vector and non-sequential inputs regarding a plurality of other customer interactions with the client, where each of the plurality of dense neural network elements comprises two dense neural networks with regularization that is bypassed with a skip connection; and an output generator configured to receive the output from the plurality of dense neural network elements and to generate a distribution of times over which a particular customer event will occur and/or a likelihood estimation that the particular customer event will occur within a particular time period.

System and method for predicting customer behavior

Various implementations of the invention store and retrieve data via a computing processor. The data is comprised of a plurality of data records with a sequentially changing data value and an other data value. For each data record in the plurality of data records, the computing processor causes the data record to be aggregated and stored with other data records in a first data storage container corresponding to a first data storage space based on the sequentially changing data value of the data record (i.e., sequential data space), and causes the data record to be aggregated and stored with other data records in a second data storage container corresponding to a second data storage space (i.e., finite data space). The data records in the second data storage container are stored based on the other data value. When a query is made against the data, the computing processor determines whether the query is best evaluated in the sequential data space or the finite data space. When the query is best evaluated in the sequential data space, the computing processor retrieves a first retrieved plurality of data records from the first data storage container. When the query is best evaluated in the finite data space, the computing processor retrieves a second retrieved plurality of data records from the second data storage container.

System and method for storing and retrieving data in different data spaces

Various implementations of the invention provide load balancing in a data storage system. A computing processor receives a request to create a new data storage container on a selected one of a plurality of data storage assets, where each of the plurality of data storage assets is configured to host a plurality of data storage containers. The computing processor determines which of the plurality of data storage assets is farthest away from an existing data storage container corresponding to the new data storage container and creates the new data storage container on the data storage asset that is farthest away from the existing data storage container.

System and method for load balancing in a data storage system

Various implementations of the invention create storage containers in a data storage system. A computing processor receives a new data record to be stored in a data storage container which is configured to store a fixed number of stored data records. The computing processor determines whether a number of the plurality of stored data records in the data storage container is within a certain threshold of the fixed number of stored data records for the data storage container. For data storage containers residing in a sequential data space, when the number of records is within the certain threshold, the computing processor opens a new data storage container, stores the new data record in the new data storage container, and closes the data storage container to new data records. For the data storage containers residing in a finite data space, when the number of records is within the certain threshold, the computing processor opens a new data storage container, splits the plurality of stored data records in the data storage container into a first plurality of stored data records and a second plurality of stored data records, where each of the first plurality of stored data records has a data value within a first range, and where each of the second plurality of stored data records has the data value within a second range. The computing processor stores the first plurality of stored data records in the data storage container, stores the second plurality of the stored data records in the new data storage container. The computing processors stores the new data record in either the data storage container or the new data storage container based on whether the data value in the new data record corresponds to the first range or the second range.

Growing community of inventors

Leesburg, VA, United States of America

Spencer Eldon Pingry