QSVM

QSVM(
    feature_map: QCallable[CArray, QNum],
    num_qubits: int,
    execution_preferences: (
        ExecutionPreferences | None
    ) = None,
)

Quantum support vector machine (QSVM) model. Classifies classical data into two categories. The model is first trained, and fitted. After pre-training, the model predicts the labels of new data points.

Parameters:

Name	Type	Description	Default
`feature_map`	`QCallable`	Feature map to be used for the kernel	required
`num_qubits`	`int`	Number of qubits used to encode each data point	required
`execution_preferences`	`ExecutionPreferences \| None`	Execution preferences for the kernel	`None`

Methods:

Name	Description
`get_qprog`	Returns the quantum program.
`get_svm_model`	Returns the classical SVM model.
`predict`	Predicts labels for new data using a precomputed kernel with a trained SVM model. Evaluates kernel
`test`	Predicts the labels of the test dataset and evaluates the resulting test score using the ground-truth labels.
`train`	Trains an SVM model using a custom precomputed kernel from the training data.

Attributes:

Name	Type	Description
`execution_preferences`
`feature_map`
`kernel_eval`
`model`
`num_qubits`
`train_data`	`ndarray \| None`

execution_preferences

execution_preferences = execution_preferences

feature_map

feature_map = feature_map

kernel_eval

kernel_eval = QuantumKernelEvaluator(
    feature_map=feature_map,
    num_qubits=num_qubits,
    execution_preferences=execution_preferences,
)

model

model = SVC(kernel='precomputed')

num_qubits

num_qubits = num_qubits

train_data

train_data: ndarray | None = None

get_qprog

get_qprog() -> QuantumProgram | None

Returns the quantum program.

get_svm_model

get_svm_model() -> SVC

Returns the classical SVM model.

predict

predict(data: ndarray) -> ndarray

Predicts labels for new data using a precomputed kernel with a trained SVM model. Evaluates kernel matrix elements which are associated with the support vectors (those associated with non-vanishing coefficients in the prediction equation).

Parameters:

Name	Type	Description	Default
`data`	`list[ndarray]`	List of new data points to predict.	required

Returns:

Type	Description
`ndarray`	np.ndarray: Predicted labels (0,1).

test

test(
    data: ndarray, data_labels: ndarray
) -> tuple[float, ndarray]

Predicts the labels of the test dataset and evaluates the resulting test score using the ground-truth labels.

Parameters:

Name	Type	Description	Default
`data`	`list[ndarray]`	List of test data points to predict.	required
`data_labels`	`ndarray`	Contains the test data labels.	required

Returns:

Name	Type	Description
`tuple`	`tuple[float, ndarray]`	containing test score (float) and test labels (np.ndarray[int]).

train

train(train_data: ndarray, train_labels: ndarray) -> None

Trains an SVM model using a custom precomputed kernel from the training data.

Parameters:

Name	Type	Description	Default
`train_data`	`ndarray`	Contains the data points (np.ndarray)	required
`train_labels`	`ndarray`	Contains the labels (0,1).	required

data_and_labels_to_data_dict

data_and_labels_to_data_dict(
    data: ndarray, labels: ndarray
) -> dict

data_dict_to_data_and_labels

data_dict_to_data_and_labels(
    data_dict: dict,
) -> tuple[ndarray, ndarray]

generate_data

generate_data(
    dimension: int = 2,
    amount: int = 20,
    amount_of_sources: int = 2,
    range_min: float = 0,
    range_max: float = 2 * pi,
    std: float = 0.5,
    sources: ndarray | None = None,
) -> dict[int, ndarray]

pauli_feature_map

pauli_feature_map(
    data: CArray[CReal],
    paulis_list: list[list[Pauli]],
    affines: list[list[float]],
    connectivity: ConnectivityType,
    reps: CInt,
    qba: QArray,
) -> None