Returns the edges between two states. Args: matrix (pd.DataFrame): A matrix attribute of the model. Returns:

Randomly change the individual's characters with probability self.mutation_rate

X: value of feature mean: the average value of feature stdev: the standard deviation of feature The function retu

Calculates the fitness of each individual in the population

Calculates the initial stationary distribution for the model. Returns: stationary (np.array): The stationary distribution.

Create children from parents by crossover

Initialize population with random strings

Input parameters: y --> array, true values y_hat --> array, predicted values Returns: float, error

Runs the Backward Algorithm. Args: input_seq (list): A list of the observed input sequence. Returns: beta (n

See the 'rules of recursion' above.

Metric to evaluate how close points in the clusters are to their centroid Returns sum of all distances of points to their corresponding cen

X: list of features Y: list consisting of target The function returns: A dictionary with Y as keys and assigned X as value

Here we calculate the value of ⲅ. For simplicity, we just calculate the denominator as a sum over all terms in the numerator, and then as

Runs the Forward Algorithm. Args: input_seq (list): A list of the observed input sequence. Returns: alpha (n

here we implement the Gaussian Density function

Calculate Eigenvalues and EigenVectors of the Laplacian Matrix. Return k eigenvectors corresponding to the smallest k eigenvalues. Use

X: list of features The function returns: A dictionary with standard deviation and mean as keys and assigned features as values.

This is the initialization step of the GMM. At this point, we must initialise our parameters μk, πk and Σk. Here we'll be using results of KMeans as

Run a single pass of K-Means X: Input data nxm matrix. n samples, m features per sample. k: Number of required clusters.

Here the value of parameters μk, πk and Σk are updated

Calculates nearest neighbor adjacency graph.

X.shape = (m,n) where, m = number of training examples n = number of features

A nicer way to print the predictions at a leaf.

Prints the model in a readable manner.

World's most elegant tree printing function.

updating weights using gradient descent

Creates a transition and emission graph of the model. Args: output_dir (str): A directory will be created with this name. If the

Runs the Viterbi Algorithm. Args: input_seq (list): A list of the observed input sequence. Returns: path (np