Aids for porting predictive models across tenants and handling impact of source changes on predictive models
AssigneeDigital.ai Software, Inc.
InventorsRahul Kapoor, Joseph Patrick Foley, Abhijeet Anant Joshi
Life of this patent
Estimating the term…
Abstract
An analytics server for scalable predictive analysis for analytics as a software service in multi-tenant environment is provided. The analytics server automatically validates portability of a predictive model from a first tenant to a second tenant by comparing value distribution of parameters between data inputs of the first tenant and the second tenant. The analytics server further automatically detects source data changes over a configurable time horizon as relevant to predictive model inputs, by comparing value distribution of parameters between two data inputs from a same tenant separated by a selected time horizon.
Claims15
A method for performing scalable predictive analysis for analytics as a software service in multi-tenant environment, the method comprising:
- obtaining a first input data associated with a first tenant;
- determining first predictive analysis parameters associated with the first tenant by analyzing the first input data for selected entities;
- comparing one or more second parameters of a second input data associated with a second tenant with the first predictive analysis parameters;
- identifying (a) a parameter from the first predictive analysis parameters that is missing in the second input data, (b) a new parameter that is present in the second input data, (c) a parameter that is present in the first predictive analysis parameters and in the second input data, and (d) a parameter that is present in the second input data and is discarded from the first predictive analysis parameters;
- comparing data associated with at least one of (i) the parameter that is present in the first predictive analysis parameters and in the second input data, and (ii) the parameter that is present in the second input data and is discarded from the first predictive analysis parameters to determine whether a data distribution of a plurality of parameters has significant changes;
- and
- porting the computer-generated first predictive model for the second tenant by (a) removing the missing parameter, or (b) incorporating (i) the new parameter that is present in the second input data on receiving a first input based on results of input data analysis of the new parameter, (ii) the parameter that is present in the first predictive analysis parameters and in the second input data and have significant changes, or (iii) the parameter that is present in the second input data and is discarded from the first predictive analysis parameters and have significant changes to obtain a transformed second predictive model for the second tenant.
A method for performing scalable predictive analysis for analytics as a software service in multi-tenant environment, the method comprising:
- obtaining a first input data associated with a first tenant;
- determining first predictive analysis parameters associated with the first tenant by analyzing the first input data for selected entities;
- determining at least one of source schema changes or data updates on the first predictive analysis parameters over a period of time by comparing the first input data with a third input data associated with the first tenant, to refine the computer-generated first predictive model based on the third input data;
- identifying at least one of (a) one or more parameters, from the first predictive analysis parameters, that are missing in the third input data, (b) one or more new parameters that are present in the third input data, (c) one or more modified parameters, or (d) one or more unchanged parameters that are present in the first input data and the third input data, based on the comparison of the first input data with the third input data, wherein the one or more new parameters, the one or more modified parameters and the one or more unchanged parameters from the third input data are analyzed using data analysis techniques, wherein data associated with at least one of the one or more modified parameters or the one or more unchanged parameters are compared to determine whether the data distribution associated with the one or more modified parameters or the one or more unchanged parameters has significant changes;
- and
- refining the computer-generated first predictive model by incorporating the source schema changes and the data updates associated with the third input to obtain a refined first predictive model for the first tenant, wherein the incorporating of the source schema changes and the data updates comprises at least one of (a) removing the one or more missing parameters from the first predictive analysis parameters, or (b) incorporating, in the computer-generated first predictive model, at least one of (i) the one or more new parameters that are present in the third input data on receiving a fourth input, (ii) the one or more modified parameters that have significant changes, or (iii) the one or more unchanged parameters that have significant changes.
Dependencies are inferred from the claim text.