API tools faq
paste
Advertisement
Guest User

Untitled

a guest
Nov 17th, 2019
108
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 10.09 KB | None | 0 0
raw download clone embed print report
  1. import numpy as np
  2. import pandas as pd
  3. from sklearn.feature_selection import SelectKBest, f_classif
  4. from sklearn.metrics import classification_report, confusion_matrix, accuracy_score
  5. from sklearn.model_selection import RepeatedStratifiedKFold
  6. from sklearn.neural_network import MLPClassifier
  7.  
  8.  
  9. def getInputAttributes():
  10. attributes = []
  11. with open("data/allhypo.names", "r") as data:
  12. for line in data.readlines():
  13. attributes.append(line)
  14. # cechy sa w linijkach od 11 do 40
  15. attributes = attributes[11:40]
  16. for index, param in enumerate(attributes):
  17. # param to jedna cecha, index to jej index w pliku
  18. attributes[index] = param.split(":")[0]
  19. return attributes
  20.  
  21.  
  22. def getInputData():
  23. thyroidData = pd.read_csv("data/allhypo.data")
  24. columns = getInputAttributes()
  25. # brakujaca nazwa kolumny z klasami
  26. columns.append('class')
  27. thyroidData.columns = columns
  28. for columnName in columns:
  29. thyroidData[columnName] = thyroidData[columnName].replace(['?'], '0')
  30. thyroidData[columnName] = thyroidData[columnName].replace(['F'], '0')
  31. thyroidData[columnName] = thyroidData[columnName].replace(['M'], '1')
  32. thyroidData[columnName] = thyroidData[columnName].replace(['t'], '1')
  33. thyroidData[columnName] = thyroidData[columnName].replace(['f'], '0')
  34. thyroidData[columnName] = thyroidData[columnName].replace(['WEST'], '0')
  35. thyroidData[columnName] = thyroidData[columnName].replace(['STMW'], '1')
  36. thyroidData[columnName] = thyroidData[columnName].replace(['SVHC'], '2')
  37. thyroidData[columnName] = thyroidData[columnName].replace(['SVI'], '3')
  38. thyroidData[columnName] = thyroidData[columnName].replace(['SVHD'], '4')
  39. thyroidData[columnName] = thyroidData[columnName].replace(['other'], '5')
  40. thyroidData.loc[thyroidData['class'].str.contains('negative', case=False), 'class'] = '0'
  41. thyroidData.loc[thyroidData['class'].str.contains('primary hypothyroid', case=False), 'class'] = '2'
  42. thyroidData.loc[thyroidData['class'].str.contains('compensated hypothyroid', case=False), 'class'] = '3'
  43. thyroidData.loc[thyroidData['class'].str.contains('secondary hypothyroid', case=False), 'class'] = '4'
  44. thyroidData.loc[thyroidData['class'].str.contains('hypothyroid', case=False), 'class'] = '1'
  45. return thyroidData
  46.  
  47.  
  48. def getParsedData(data, ranking, numberOfAttributes):
  49. attributesToConsider = []
  50. parsedData = data
  51. for attribute in range(numberOfAttributes):
  52. attributesToConsider.append(ranking[attribute].get("attribute"))
  53. for column in data:
  54. if (not column in attributesToConsider):
  55. parsedData = parsedData.drop(column, axis=1)
  56. return parsedData
  57.  
  58.  
  59. def runNeutralNetworkClassifer(X_train, X_test, y_train, y_test, perceptronCount):
  60. mlpc = MLPClassifier(hidden_layer_sizes=perceptronCount, max_iter=500)
  61. mlpc.fit(X_train, y_train)
  62. pred_mlpc = mlpc.predict(X_test)
  63. print(classification_report(y_test, pred_mlpc))
  64. print(confusion_matrix(y_test, pred_mlpc))
  65. print(accuracy_score(y_test, pred_mlpc))
  66.  
  67.  
  68. def runNestedCrossValidation(X, y, numberOfTrials):
  69. rskf = RepeatedStratifiedKFold(2, numberOfTrials, random_state=123)
  70. mlpScores = []
  71. confusionMatrix = []
  72. bestScore = 0
  73. it = 1
  74. for train_index, test_index in rskf.split(X, y):
  75. it = it + 1
  76. X_train, X_test = X.iloc[train_index], X.iloc[test_index]
  77. y_train, y_test = y.iloc[train_index], y.iloc[test_index]
  78. mlpc = MLPClassifier(hidden_layer_sizes=11, max_iter=2000)
  79. mlpc.fit(X_train, y_train)
  80. pred_mlpc = mlpc.predict(X_test)
  81. score = accuracy_score(y_test, pred_mlpc)
  82. mlpScores.append(score)
  83. if score > bestScore:
  84. bestScore = score
  85. confusionMatrix = confusion_matrix(y_test, pred_mlpc)
  86.  
  87. return np.mean(mlpScores), confusionMatrix
  88.  
  89.  
  90. def runBackpropagation(X, y, perceptronCount, momentum_, numberOfTrials):
  91. rskf = RepeatedStratifiedKFold(2, numberOfTrials, random_state=123)
  92. bpScores = []
  93. confusionMatrix = []
  94. bestScore = 0
  95. for train_index, test_index in rskf.split(X, y):
  96. X_train, X_test = X.iloc[train_index], X.iloc[test_index]
  97. y_train, y_test = y.iloc[train_index], y.iloc[test_index]
  98. nn = MLPClassifier(activation='logistic', solver='sgd', hidden_layer_sizes=perceptronCount, max_iter=2000,
  99. random_state=1,
  100. momentum=momentum_)
  101. nn.fit(X_train, y_train)
  102. pred = nn.predict(X_test)
  103. score = accuracy_score(y_test, pred)
  104. bpScores.append(score)
  105. if score > bestScore:
  106. bestScore = score
  107. confusionMatrix = confusion_matrix(y_test, pred)
  108.  
  109. return np.mean(bpScores), confusionMatrix
  110.  
  111.  
  112. def runAllExperiments(X_, y, features):
  113. numberOfNeuronsArray = [11, 17, 23]
  114. numberOfTrials = 5
  115. bestResult = 0
  116. confusion_matrix = []
  117. f = open("experimentResults.txt", "w+")
  118. f.write("-----WYNIKI EKSPERYMENTOW DLA SIECI NEURONOWEJ DLA PRZYPADKU NIEDOCZYNNOSCI TARCZYCY-----\n")
  119. print("-----WYNIKI EKSPERYMENTOW DLA SIECI NEURONOWEJ DLA PRZYPADKU NIEDOCZYNNOSCI TARCZYCY-----")
  120. for it in range(1, 8):
  121. X = getParsedData(X_, features, it)
  122. f.write(str(it) + " Wyniki eksperymentow dla " + str(it) + " cech(y)\n")
  123. print(str(it) + " Wyniki eksperymentow dla " + str(it) + " cech(y)")
  124. for j in range(0, 3):
  125. numberOfNeurons = numberOfNeuronsArray[j]
  126. f.write(str(it) + "." + str(j + 1) + " Wyniki dla liczby neuronow w warstwie ukrytej rownej " + str(
  127. numberOfNeurons) + "\n")
  128. f.write("Siec jednokierunkowa: \n" )
  129. print(str(it) + "." + str(j + 1) + " Wyniki dla liczby neuronow w warstwie ukrytej rownej " + str(
  130. numberOfNeurons))
  131. print("Siec jednokierunkowa: ")
  132. result, confMatrix = runNestedCrossValidation(X, y, numberOfTrials)
  133. f.write(str(result) + "\n")
  134. print(str(result))
  135. if (result > bestResult):
  136. bestResult = result
  137. confusionMatrix = confMatrix
  138.  
  139. for k in range(0, 2):
  140. f.write("Propagacja wsteczna z momentum rownym " + str(k) + ": \n")
  141. print("Propagacja wsteczna z momentum rownym " + str(k) + ": ")
  142. result, confMatrix = runBackpropagation(X, y, numberOfNeurons, k, numberOfTrials)
  143. f.write(str(result) + "\n")
  144. print(result)
  145. if (result > bestResult):
  146. bestResult = result
  147. confusionMatrix = confMatrix
  148. f.write("Najlepszy uzyskany wynik: \n")
  149. f.write(str(bestResult) + "\n")
  150. print("Najlepszy uzyskany wynik: ")
  151. print(bestResult)
  152. f.write("Najlepsza macierz konfuzji: \n")
  153. f.write(confusionMatrix)
  154. print("Najlepsza macierz konfuzji: ")
  155. print(confusionMatrix)
  156. f.close()
  157.  
  158. def runExperiments(X_, y, features, numberOfFeatures):
  159. numberOfNeuronsArray = [11, 17, 23]
  160. numberOfTrials = 5
  161. bestResult = 0
  162. confusionMatrix = []
  163. f = open("experimentResults.txt", "w+")
  164. f.write("-----WYNIKI EKSPERYMENTOW DLA SIECI NEURONOWEJ DLA PRZYPADKU NIEDOCZYNNOSCI TARCZYCY-----\n")
  165. print("-----WYNIKI EKSPERYMENTOW DLA SIECI NEURONOWEJ DLA PRZYPADKU NIEDOCZYNNOSCI TARCZYCY-----")
  166.  
  167. X = getParsedData(X_, features, numberOfFeatures)
  168. f.write("Wyniki eksperymentow dla " + str(numberOfFeatures) + " cech\n")
  169. print("Wyniki eksperymentow dla " + str(numberOfFeatures) + " cech")
  170. for j in range(0, 3):
  171. numberOfNeurons = numberOfNeuronsArray[j]
  172. f.write("Wyniki dla liczby neuronow w warstwie ukrytej rownej " + str(numberOfNeurons) + "\n")
  173. f.write("Siec jednokierunkowa: \n")
  174. print("Wyniki dla liczby neuronow w warstwie ukrytej rownej " + str(numberOfNeurons))
  175. print("Siec jednokierunkowa: ")
  176. result, confMatrix = runNestedCrossValidation(X, y, numberOfTrials)
  177. f.write(str(result) + "\n")
  178. print(str(result))
  179. if (result > bestResult):
  180. bestResult = result
  181. confusionMatrix = confMatrix
  182.  
  183. for k in range(0, 2):
  184. f.write("Propagacja wsteczna z momentum rownym " + str(k) + ": \n")
  185. print("Propagacja wsteczna z momentum rownym " + str(k) + ": ")
  186. result, confMatrix = runBackpropagation(X, y, numberOfNeurons, k, numberOfTrials)
  187. f.write(str(result) + "\n")
  188. print(result)
  189. if (result > bestResult):
  190. bestResult = result
  191. confusionMatrix = confMatrix
  192.  
  193. f.write("Najlepszy uzyskany wynik: \n")
  194. f.write(str(bestResult) + "\n")
  195. print("Najlepszy uzyskany wynik: ")
  196. print(bestResult)
  197. f.write("Najlepsza macierz konfuzji: \n")
  198. f.write(str(confusionMatrix))
  199. print("Najlepsza macierz konfuzji: ")
  200. print(str(confusionMatrix))
  201. f.close()
  202.  
  203.  
  204. def main():
  205. # funkcja zwraca macierz ilosci probek i wartosci cech
  206. objects = getInputData()
  207. columns = objects.columns
  208. # znajduje stałą w 26 i 27. do usuniecia
  209. to_delete = [columns[26], columns[27]]
  210. objects = objects.drop(to_delete[0], axis=1)
  211. objects = objects.drop(to_delete[1], axis=1)
  212. # x co techy i wartosci
  213. X = objects.drop('class', axis=1)
  214. # y to tylko wartosci class
  215. y = objects['class']
  216. # ilosc cech
  217. new_features = X.shape[1]
  218. k_best_selector = SelectKBest(score_func=f_classif, k=new_features)
  219. k_best_selector.fit(X, y)
  220. scores = k_best_selector.scores_
  221. column_scores = [{'attribute': name, 'result': round(score, 2)} for name, score in zip(X.columns, scores)]
  222. features = sorted(column_scores, key=lambda x: x['result'], reverse=True)
  223. print("Ranking cech:")
  224. for row in features:
  225. print(row)
  226. numberOfAttributes = 7 # przy 6 przestaje rosnąć
  227. numberOfNeurons = 11
  228. numberOfTrials = 5
  229.  
  230. runExperiments(X, y, features, 6)
  231.  
  232.  
  233. if __name__ == "__main__":
  234. main()
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement
Public Pastes
Advertisement
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!