Generative modeling methods and systems for simulating sensor measurements

A method of estimating a glucose level using a glucose sensing arrangement, the method comprising:

• obtaining a generative model associated with the glucose sensing arrangement;
• obtaining reference glucose measurement data from a blood glucose meter;
• applying the generative model to the reference glucose measurement data to obtain simulated measurement data for the glucose sensing arrangement;
• determining an estimation model for the glucose sensing arrangement based at least in part on relationships between the simulated measurement data and the reference glucose measurement data;
• and
• applying the estimation model to one or more output measurements from a sensing element of the glucose sensing arrangement to obtain an estimated value for the glucose level using the glucose sensing arrangement.

A system comprising:

• a database to store historical calibration data corresponding to instances of a first sensing arrangement, the historical calibration data comprising respective sets of a reference measurement value, a calibration factor value, and a timestamp value;
• and
• a server coupled to the database and a network to obtain the historical calibration data from the database, determine a generative model based on relationships between measurement data obtained from a second sensing arrangement and temporally associated calibration data, apply the generative model associated with the second sensing arrangement to the historical calibration data to obtain respective sets of simulated measurement data for the second sensing arrangement as a function of the respective sets of the calibration factor value and the timestamp value, determine an estimation model for the second sensing arrangement based at least in part on relationships between the respective sets of simulated measurement data and the corresponding respective reference measurement values, and provide the estimation model associated with the second sensing arrangement to a computing device via the network, wherein the computing device applies the estimation model to one or more output measurements from a sensing element of an instance of the second sensing arrangement to obtain an estimated calibrated measurement value.