final_pred = []
for i in range(len(list_Xt)):
#xaxis.append(i)
print("Step " + str(i))
X = list_Xt[i]
y = list_yt[i]
y_cls = y
outlier_indices = det.detect_outliers(X)
#outlier_report(y_cls.values, outlier_indices)
out_idx.extend([x + breakpoints[i] for x in outlier_indices])
normal_indices = X.index.difference(outlier_indices)
det.update_outlier(X.iloc[normal_indices])
outlier_X = X.iloc[outlier_indices]
y_pred = det.classfiy(outlier_X)
outlier_y = y_cls.iloc[outlier_indices]
print("")
#cls_report.append(classification_report(outlier_y.values, y_pred, output_dict=True, labels=classes))
#print(classification_report(outlier_y.values, y_pred, labels=classes))
print(outlier_y.values.shape)
correct_indices, wrong_indices = classification_result(
outlier_y.values, y_pred)
correct_X = outlier_X.iloc[correct_indices]
while i < size:
print(str(i) + "/" + str(size))
j = rd.randint(i + MIN_SIZE, i + MAX_SIZE)
j = min(j, size)
X = pd.DataFrame(df_X[i:j].values,
columns=new_features,
dtype=np.float64)
y = np.ndarray.flatten(
pd.DataFrame(df_y[i:j].values, columns=['class'],
dtype=np.int64).values)
if i == 0:
det.initialize(X, X, y)
i = j
continue
det.update_outlier(X)
#det.update_classifier(X, y)
i = j
print("Training Complete")
cls_reports = []
breakpoints = {x: [] for x in file_list[1:]}
for file in file_list[1:]:
print("Loading data...")
df_X, df_y = read_file(file)
df_X = pd.DataFrame(df_X, columns=new_features, dtype=np.float64)
print("Done")
print("Analyzing file " + file)
out_idx = []
final_pred = []