Naive Bayes¶

def predict(self, sample: list[float]) -> int:
    """Predict the class label for a given sample.

    Args:
        sample: A single sample to predict, represented as a list of feature values.

    Returns:
        int: Predicted class label.
    """
    class_posteriors = self.predict_proba(sample)
    label = max(class_posteriors, key=lambda label: class_posteriors[label])
    return label

def predict_proba(self, sample: list[float], normalization: bool = True) -> dict[int, float]:
    """Predict class probabilities for a given sample.

    Args:
        sample: A single sample to predict, represented as a list of feature values.
        normalization: Whether to normalize the probabilities. If False, return
            joint probabilities (or densities for Gaussian features). Default is True.

    Returns:
        dict[int, float]: Posterior probabilities by class label when normalized;
            otherwise, unnormalized joint probabilities or densities.
    """
    class_log_probs = self.predict_log_proba(sample, normalization)
    return {label: math.exp(log_prob) for label, log_prob in class_log_probs.items()}

def predict_log_proba(self, sample: list[float], normalization: bool = True) -> dict[int, float]:
    """Predict log probabilities for a given sample.

    Args:
        sample: A single sample to predict, represented as a list of feature values.
        normalization: Whether to normalize the log probabilities. If False, return
            joint log probabilities (or log densities for Gaussian features). Default is True.

    Returns:
        dict[int, float]: Log posterior probabilities by class label when normalized;
            otherwise, unnormalized joint log probabilities or log densities.
    """
    if not self.class_prior_:
        msg = "The model is not fitted yet!"
        raise ValueError(msg)
    class_log_likelihoods = self._log_likelihood(sample)
    class_log_joint_probs: dict[int, float] = {}
    for label, log_likelihood in class_log_likelihoods.items():
        class_log_joint_probs[label] = log_likelihood + math.log(self.class_prior_[label])
    if normalization is False:
        return class_log_joint_probs
    # ref: https://github.com/scikit-learn/scikit-learn/blob/2beed55847ee70d363bdbfe14ee4401438fba057/sklearn/naive_bayes.py#L97
    max_log_joint_prob = max(class_log_joint_probs.values())
    log_evidence = max_log_joint_prob + math.log(
        sum(math.exp(log_joint_prob - max_log_joint_prob) for log_joint_prob in class_log_joint_probs.values()),
    )
    class_log_posteriors = {
        label: log_joint_prob - log_evidence for label, log_joint_prob in class_log_joint_probs.items()
    }
    return class_log_posteriors

def fit(self, dataset: list[list[float]], labels: list[int]) -> GaussianNaiveBayes:
    """Fit the Gaussian Naive Bayes classifier.

    Args:
        dataset: Training data, where each row is a sample and each column is a feature.
        labels: Target labels for training data.

    Returns:
        self: Returns the instance itself.
    """
    self.labels_ = sorted(set(labels))
    self.class_count_ = len(set(labels))
    self.class_prior_ = {label: count / len(dataset) for label, count in Counter(labels).items()}
    self.epsilon_ = self.var_smoothing * max(self._variance(col) for col in zip(*dataset, strict=False))
    self.means_, self.variances_ = self._get_classes_means_variances(dataset, labels)
    return self

def fit(self, dataset: list[list[float]], labels: list[int]) -> MultinomialNaiveBayes:
    """Fit the Multinomial Naive Bayes classifier.

    Args:
        dataset: Training data, where each row is a sample and each column is a feature.
                 Features should be represented as counts (non-negative integers).
        labels: Target labels for training data.

    Returns:
        self: Returns the instance itself.
    """
    self.labels_ = sorted(set(labels))
    self.class_count_ = len(set(labels))
    # get the prior from training dataset labels
    self.class_prior_ = {label: count / len(dataset) for label, count in Counter(labels).items()}
    self.class_feature_count_, self.class_feature_log_prob_ = self._get_classes_feature_counts_log_probs(
        dataset, labels
    )
    return self

def fit(self, dataset: list[list[float]], labels: list[int]) -> CategoricalNaiveBayes:
    """Fit the Categorical Naive Bayes classifier.

    Args:
        dataset: Training data, where each row is a sample and each column is a feature.
        labels: Target labels for training data.

    Returns:
        self: Returns the instance itself.
    """
    self.labels_ = sorted(set(labels))
    self.class_count_ = len(set(labels))
    # get the prior from training dataset labels
    self.class_prior_ = {label: count / len(dataset) for label, count in Counter(labels).items()}
    self.class_feature_count_, self.class_feature_log_prob_ = self._get_classes_feature_counts_log_probs(
        dataset, labels
    )
    return self