def fit()

in graspologic/cluster/kclust.py [0:0]

• 31 lines of code
• 6 McCabe index (conditional complexity)

    def fit(self, X: np.ndarray, y: Optional[np.ndarray] = None) -> "KMeansCluster":
        """
        Fits kmeans model to the data.

        Parameters
        ----------
        X : array-like, shape (n_samples, n_features)
            List of n_features-dimensional data points. Each row
            corresponds to a single data point.

        y : array-like, shape (n_samples,), optional (default=None)
            List of labels for `X` if available. Used to compute ARI scores.

        Returns
        -------
        self
        """
        # Deal with number of clusters
        if self.max_clusters > X.shape[0]:
            msg = "n_components must be >= n_samples, but got \
                n_components = {}, n_samples = {}".format(
                self.max_clusters, X.shape[0]
            )
            raise ValueError(msg)
        else:
            max_clusters = self.max_clusters

        # Get parameters
        random_state = self.random_state

        # Compute all models
        models = []
        silhouettes = []
        aris = []
        for n in range(2, max_clusters + 1):
            model = KMeans(n_clusters=n, random_state=random_state)

            # Fit and compute values
            predictions = model.fit_predict(X)
            models.append(model)
            silhouettes.append(silhouette_score(X, predictions))
            if y is not None:
                aris.append(adjusted_rand_score(y, predictions))

        if y is not None:
            self.ari_ = aris
            self.silhouette_ = silhouettes
            self.n_clusters_ = np.argmax(aris) + 1
            self.model_ = models[np.argmax(aris)]
        else:
            self.ari_ = None
            self.silhouette_ = silhouettes
            self.n_clusters_ = np.argmax(silhouettes) + 1
            self.model_ = models[np.argmax(silhouettes)]

        return self