def fit(self, X, y=None): if self._empty_fit(): return self if self.dtype is not None: X = cast(X, self.dtype) mask = self._missing_value_mask(X) X = numpy.ma.masked_array(X, mask=mask) min = numpy.asarray(numpy.nanmin(X, axis=0)) max = numpy.asarray(numpy.nanmax(X, axis=0)) if self.with_data: self.data_min_ = min self.data_max_ = max if self.with_statistics: self.counts_ = _count(mask) X = numpy.ma.asarray(X, dtype=numpy.float).filled(float("NaN")) self.numeric_info_ = { "minimum": min, "maximum": max, "mean": numpy.asarray(numpy.nanmean(X, axis=0)), "standardDeviation": numpy.asarray(numpy.nanstd(X, axis=0)), "median": numpy.asarray(numpy.nanmedian(X, axis=0)), "interQuartileRange": numpy.asarray(_interquartile_range(X, axis=0)) } return self
def transform(self, X): if self.dtype is not None: X = cast(X, self.dtype) missing_value_mask = self._missing_value_mask(X) nonmissing_value_mask = ~missing_value_mask valid_value_mask = self._valid_value_mask(X, nonmissing_value_mask) invalid_value_mask = ~numpy.logical_or(missing_value_mask, valid_value_mask) self._transform_missing_values(X, missing_value_mask) self._transform_valid_values(X, valid_value_mask) self._transform_invalid_values(X, invalid_value_mask) return X
def fit(self, X, y = None): X = column_or_1d(X, warn = True) if self._empty_fit(): return self if self.dtype is not None: X = cast(X, self.dtype) mask = self._missing_value_mask(X) values, counts = numpy.unique(X[~mask], return_counts = True) if self.with_data: if (self.missing_value_replacement is not None) and numpy.any(mask) > 0: self.data_ = numpy.unique(numpy.append(values, self.missing_value_replacement)) else: self.data_ = values if self.with_statistics: self.counts_ = _count(mask) self.discr_stats_ = (values, counts) return self
def transform(self, X): if hasattr(self, "dtype"): X = cast(X, self.dtype) if self.outlier_treatment == "as_missing_values": mask = self._outlier_mask(X) if hasattr(self, "missing_values"): if type(self.missing_values) is list: raise ValueError() X[mask] = self.missing_values else: X[mask] = None elif self.outlier_treatment == "as_extreme_values": mask = self._negative_outlier_mask(X) X[mask] = self.low_value mask = self._positive_outlier_mask(X) X[mask] = self.high_value return super(ContinuousDomain, self).transform(X)
def transform(self, X): return cast(X, self.dtype)
def transform(self, X): func = lambda x: self._eval_row(x) Xt = eval_rows(X, func) if hasattr(self, "dtype"): Xt = cast(Xt, self.dtype) return _col2d(Xt)
def transform(self, X): func = lambda x: self._eval_row(x) Xt = eval_rows(X, func) if self.dtype is not None: Xt = cast(Xt, self.dtype) return _col2d(Xt)
def transform(self, X): if hasattr(self, "dtype"): X = cast(X, self.dtype) return super(TemporalDomain, self).transform(X)