Mixture models (C API)

typedef struct _EmlMixtureModel EmlMixtureModel

Model.

Normally initialized by code generated using emlearn.

int32_t eml_mixture_score(EmlMixtureModel *model, const float values[], int32_t values_length, float *probabilities, float *out_score)

Run inference and return component-wise probabilities and outlier score

Parameters:

• model – Model instance

• values – Input data values

• values_length – Input data values

• probabilities – Return location for probabilities. Must be length == n_components

• out_score – Return location for the overall outlier score (single value).

Returns:

EmlOk on success, or error on failure

int32_t eml_mixture_log_proba(EmlMixtureModel *model, const float values[], int32_t values_length, float *probabilities)

Run inference and return log-probability scores.

Low-level function that underlies eml_mixture_score() Useful when operating with log probabilities is preferred.

Parameters:

• model – Model instance

• values – Input data values

• values_length – Input data values

• probabilities – Return location for probabilities. Must be length == n_components

Returns:

EmlOk on success, or error on failure