change default to "Normalize=False" and remove "normalize_type: 'linear"

[jkoschinsky]

Two updates:

• change default from normalize=True (bold below) Normalize=False
• for normalize_type: 'linear' or 'z-score', remove 'linear' since it doesn't work well with all the zero values (not enough walkable values)

These are the specification of the parameters for AccessModel:

def calculate(self,upper_threshold,category_weight_dict=None, normalize=True, normalize_type='linear')

normalize: boolean. If true, results will be normalized from 0 to 100. normalize_type: 'linear' or 'z_score'.

def _normalize(self, column, normalize_type):
    """
    Normalize results.
    Args:
        column: which column to normalize.
        normalize_type: 'linear' or 'z-score'

    Raises:
        UnexpectedEmptyColumnException
        UnexpectedNormalizeTypeException
    """
    if normalize_type == 'linear':
        max_score = self.model_results[column].max()
        self.model_results[column] = (self.model_results[column] / max_score) * 100.0
    elif normalize_type == 'z_score':
        try:
            self.model_results[column] = (self.model_results[column]
                                          - self.model_results[column].mean()) / self.model_results[column].std()