Abstract: Implementations relate to methods, systems, and computer-readable media to perform customized matchmaking.

Abstract: Implementations relate to methods, systems, and computer-readable media to perform customized matchmaking. In some implementations, the method may include receiving a request from a user device of a requesting user to join a particular virtual experience, identifying virtual experience instances that are in progress, determining a first set of virtual experience instances eligible for the requesting user to join, obtaining one or more matchmaking rules that are associated with the particular virtual experience, determining a second set of virtual experience instances that satisfy the one or more matchmaking rules, determining a respective score for individual virtual experience instances in the second set of virtual experience instances based on one or more matchmaking scoring criteria, determining, based on the respective scores, a third set of virtual experience instances, and causing the user device to connect to the server that hosts a particular one of the third set of virtual experience instances.

Abstract: Forecasts are provided based on dynamic model selection for different sets of time series. A model comprises a transformation and a prediction algorithm. Given a time series, a transformation is selected for the time series and a prediction algorithm is selected to make a forecast based on the transformed time series. Sets of time series are distinguished from each other based on diverse sparsities, temporal scales and other time series attributes. A model is dynamically selected based on time series attributes to increase forecasting accuracy and decrease forecasting computation time. The dynamic model selection is based on the creation of a meta-model from historical sets of historical time series.

Abstract: Forecasts are provided based on dynamic model selection for different sets of time series. A model comprises a transformation and a prediction algorithm. Given a time series, a transformation is selected for the time series and a prediction algorithm is selected to make a forecast based on the transformed time series. Sets of time series are distinguished from each other based on diverse sparsities, temporal scales and other time series attributes. A model is dynamically selected based on time series attributes to increase forecasting accuracy and decrease forecasting computation time. The dynamic model selection is based on the creation of a meta-model from historical sets of historical time series.

Abstract: A method for cash forecasting creates a plurality of event streams associated with a user of an online accounting service, identifies a first pattern in a first event stream that is a commitment and a second pattern in a second event stream that is a repeated pattern, generates a characterization for each of the plurality of event streams based on each identified pattern for the event stream, predicts one or more expected cash payments and one or more corresponding expected dates for each of the event streams using the characterization for the event stream and a forecasting model, receives a query related to cash flow through an application program interface (API), and responds to the query based on the one or more of the expected cash payments, the corresponding expected dates, and the one or more measures of uncertainty, to the query.