Methods of characterizing similarity between measurements on entities, computer programs product and data carrier

A method for characterizing the similarity between measurements of a plurality of entities (i) comprising a first entity (i′) and a second entity (i″), comprising:

• taking measurements (wi,j) at a plurality of measurement points (j) per entity (i) using a sensor;
• defining a model, the model comprising a stochastic process and a model function (h), the model function (h) having values which depend on a set of parameters (β) and the measurement points (j);
• estimating the set of parameters (β) by fitting the model function to the measurements (wi,j);
• determining residual data (Zi,j) at the plurality of measurement points (j) for the first entity (i′) and the second entity (i″) by subtracting the fitted model function from the measurements;
• estimating a correlation coefficient (ri′,i″) for the first entity (i′) and the second entity (i″) based on the determined residual data (Zi,j);
• and
• using the estimated correlation coefficient (ri′,i″) to characterize the similarity between the measurements;
• wherein the model is defined such that the residual data is expected to have a deterministic component which depends on the measurement points and that dominates the estimate of the correlation coefficient and estimating the correlation coefficient (ri′,i″) using an estimate for the entity average residue averaged over the measurement points of the first entity (i′) and using an estimate for the entity average residue averaged over the measurement points of the second entity (i″).

A computer program product comprising instructions and data stored on a non-transitory computer readable medium and which when executed by a computer processor, causes the computer to characterize the similarity between measurements of a plurality of entities (i) comprising a first entity (i′) and a second entity (i″), by a process comprising:

• receiving measurements (wi,j) taken at a plurality of measurement points (j) per entity (i);
• defining a model comprising a stochastic process and a model function (h), the model function (h) having values which depend on a set of parameters (β) and the measurement points (j);
• estimating the set of parameters (β) by fitting the model function to the measurements (wi,j);
• determining residual data (Zi,j) at the plurality of measurement points (j) for the first entity (i′) and the second entity (i″) by subtracting the fitted model function from the measurements;
• estimating a correlation coefficient (ri′,i″) for the first entity (i′) and the second entity (i″) based on the determined residual data (Zi,j);
• using the estimated correlation coefficient (ri′,i″) to characterize the similarity between the measurements;
• wherein the model is defined such that the residual data is expected to have a deterministic component which depends on the measurement points and that dominates the estimate of the correlation coefficient and estimating the correlation coefficient (ri′,i″) using an estimate for the entity average residue averaged over the measurement points of the first entity (i′) and using an estimate for the entity average residue averaged over the measurement points of the second entity (i″).