fdars.classification¶

Classification methods for functional data via FPC score projection.

Functions¶

Function	Description
`fclassif_lda`	Linear discriminant analysis
`fclassif_qda`	Quadratic discriminant analysis
`fclassif_knn`	K-nearest neighbors classification
`fclassif_kernel`	Kernel classification
`fclassif_cv`	Cross-validated classification

fdars.fclassif_lda(data, labels, ncomp=3)

LDA classification for functional data. Projects onto FPC scores, then applies linear discriminant analysis.

Parameter	Type	Default	Description
`data`	`ndarray (n, m)`		Functional data
`labels`	`ndarray (n,)` of `int64`		Class labels
`ncomp`	`int`	`3`	Number of FPC components

Returns	Type	Description
result	`dict`	Keys: `predicted` (n,), `accuracy`

result = fdars.fclassif_lda(data, labels, ncomp=5)
print(f"Accuracy: {result['accuracy']:.3f}")

fdars.fclassif_qda(data, labels, ncomp=3)

QDA classification for functional data. Uses class-specific covariance matrices.

Parameter	Type	Default	Description
`data`	`ndarray (n, m)`		Functional data
`labels`	`ndarray (n,)` of `int64`		Class labels
`ncomp`	`int`	`3`	Number of FPC components

Returns	Type	Description
result	`dict`	Keys: `predicted` (n,), `accuracy`

result = fdars.fclassif_qda(data, labels, ncomp=5)

fdars.fclassif_knn(data, labels, ncomp=3, k=5)

K-nearest neighbors classification in FPC score space.

Parameter	Type	Default	Description
`data`	`ndarray (n, m)`		Functional data
`labels`	`ndarray (n,)` of `int64`		Class labels
`ncomp`	`int`	`3`	Number of FPC components
`k`	`int`	`5`	Number of nearest neighbors

Returns	Type	Description
result	`dict`	Keys: `predicted` (n,), `accuracy`

result = fdars.fclassif_knn(data, labels, ncomp=5, k=7)

fdars.fclassif_kernel(data, argvals, labels, h_func=1.0, h_scalar=1.0)

Kernel classification directly on functional data using functional and scalar bandwidths.

Parameter	Type	Default	Description
`data`	`ndarray (n, m)`		Functional data
`argvals`	`ndarray (m,)`		Evaluation points
`labels`	`ndarray (n,)` of `int64`		Class labels
`h_func`	`float`	`1.0`	Functional bandwidth
`h_scalar`	`float`	`1.0`	Scalar bandwidth

Returns	Type	Description
result	`dict`	Keys: `predicted` (n,), `accuracy`

result = fdars.fclassif_kernel(data, t, labels, h_func=0.5, h_scalar=0.5)

fdars.fclassif_cv(data, argvals, labels, method="lda", ncomp=3, nfold=5)

Cross-validated classification with error rate estimation.

Parameter	Type	Default	Description
`data`	`ndarray (n, m)`		Functional data
`argvals`	`ndarray (m,)`		Evaluation points
`labels`	`ndarray (n,)` of `int64`		Class labels
`method`	`str`	`"lda"`	`"lda"`, `"qda"`, `"knn"`, or `"kernel"`
`ncomp`	`int`	`3`	Number of FPC components
`nfold`	`int`	`5`	Number of CV folds

Returns	Type	Description
result	`dict`	Keys: `error_rate`, `fold_errors` (nfold,), `best_ncomp`

result = fdars.fclassif_cv(data, t, labels, method="knn", ncomp=5, nfold=10)
print(f"CV error rate: {result['error_rate']:.3f}")