Skip to content

Math

Matrix

ALG_Matrix

Type : Structure

Description : A matrix.

Variable Type Accessibility Description
values double Read/Write A 1D array that contains all the matrix's values.
_size size_t Read Only An array that describe matrix size. First value is the number of rows. The second value is the number of columns

ALG_CreateMatrix()

Type : Function

Prototype : ALG_Matrix *ALG_CreateMatrix(size_t rows, size_t columns, double v)

Description : Create a matrix.

Parameters

Parameter Type Description
rows int The number of rows.
columns int The number of columns.
v double The value to init all the matrix.

Mutations and return

Case Name Type Description
Return ALG_Matrix* A pointer to the matrix structure that has been created.

ALG_FillMatrix()

Type : Function

Prototype : void ALG_FillMatrix(ALG_Matrix *m, double v)

Description : Fill all the matrix m with the value v.

Parameters

Parameter Type Description
m ALG_Matrix* The matrix to fill.
v double The value to use for filling the matrix.

Mutations and return

Case Name Type Description
Mutation m ALG_Matrix* All the values of the matrix are set to v.

ALG_DebugMatrix()

Type : Function

Prototype : void ALG_DebugMatrix(ALG_Matrix *m)

Description : Print information about the matrix in the console.

Parameters

Parameter Type Description
m ALG_Matrix* The matrix to debug.

Mutations and return

No mutations or return.


ALG_DestroyMatrix()

Type : Function

Prototype : void ALG_DestroyMatrix(ALG_Matrix *m)

Description : Free memory of the matrix m.

Parameters

Parameter Type Description
m ALG_Matrix* The matrix to free.

Mutations and return

Case Name Type Description
Mutation m ALG_Matrix* Free the pointer.

Activation


ALG_Sigmoid()

Type : Function

Prototype : double ALG_Sigmoid(double x, double lambda)

Description : Compute the sigmoid function with param lambda. The function is computed with the formula :

$$ f_{\lambda}(x) = \frac{1}{1 + e^{- \lambda x}} $$

Parameters

Parameter Type Description
x double The point where sigmoid must be evaluated.
lambda double The sigmoid parameter.

Mutations and return

Case Name Type Description
Return double The computed value

Cost functions


ALG_MeanSquaredError()

Type : Function

Prototype : double ALG_MeanSquaredError(double predictions[], double targets[], size_t batchSize)

Description : Compute the mean squared function. It is computed with the formula :

$$ MSE(x) = \frac{1}{n} \sum_{i = 1}^{n}(Y_i - \hat{Y}_i)^2 $$

where $n$ is the number of predictions, $Y_i$ is the target prediction et $\hat{Y}_i$ is the predicted value.

Parameters

Parameter Type Description
predictions double[] An array of predicted values. Must have the same size as targets
targets double[] An array of targets values. Must have the same size as predictions
batchSize size_t[] The number of predictions.

Mutations and return

Case Name Type Description
Return double The computed value