Merge pull request #2531 from taneemishere/Taneem_Jan

All machine learning algorithms are being moved to machine_learning directory plus SVM in python is also added.

machine_learning/Support_Vector_Machine/Python/README.md

@@ -0,0 +1,8 @@
+# Supprt Vector Machine
+
+The Supprt Vector Machine is one of the superivsed learning algorithm of machine learning that is 
+used for both the classification and the regression problems. The main classification of the data 
+points are done through by drawing the optimal hyperplane. But how would the hyperplane be determine
+as the optimal one. Well this algorithm does this drawing the supporting vetors the categories in 
+the dataset. And the main hyperplane would be consider the optimal one that has the wider area between 
+supporting vector.

machine_learning/Support_Vector_Machine/Python/SVM_with_Sklearn.py

@@ -0,0 +1,40 @@
+# some imports 
+
+from sklearn import datasets
+from sklearn.model_selection import train_test_split
+from sklearn import svm
+from sklearn import metrics
+
+# read the dataset from sklearn dataset
+cancer = datasets.load_breast_cancer()
+
+# See the features and label names of the dataset
+print("Features are: ", cancer.feature_names)
+print("Labels are: ", cancer.target_names)
+
+# Assign the values to X as featrues and to y the labels
+X = cancer.data
+y = cancer.target
+
+# Split the dataset into 80% and 20% for training and testing respectively
+x_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
+
+# print(x_train, y_train)
+
+# these are the two classes of the label
+classes = ['malignant', 'benign']
+
+# Support Vector Classifier of Support Vector Machine
+# Here the C is the Soft Margin for the SVM
+
+clf = svm.SVC(kernel="linear", C=2)
+clf.fit(x_train, y_train)
+
+# predict the values of training features
+y_pred = clf.predict(x_test)
+
+# seeing the acuuracy score of the model
+acc = metrics.accuracy_score(y_test, y_pred)
+
+print("Accuracy of SVC: ", acc)
+

deep_learning/python/transfer-learning/readme.md → machine_learning/deep_learning/python/transfer-learning/readme.md

@@ -1,11 +1,11 @@
-# Transfer Learning
-In Transfer Learning, the knowledge of an already trained Machine Learning model is applied to a different but related problem. For
-example, if you trained a simple classifier to predict whether an image contains a backpack, you could use the knowledge that the 
-model gained during its training to recognize other objects like sunglasses. With transfer learning, we basically try to exploit 
-what has been learned in one task to improve generalization in another. We transfer the weights that a Network has learned at Task 
-A to a new Task B.<br>
-The general idea is to use knowledge, that a model has learned from a task where a lot of labeled training data is available, in a 
-new task where we don’t have a lot of data. Instead of starting the learning process from scratch, you start from patterns that 
-have been learned from solving a related task. Transfer Learning is mostly used in Computer Vision and Natural Language Processing 
-Tasks like Sentiment Analysis, because of the huge amount of computational power that is needed for them.
-![alt workflow](https://github.com/BAJUKA/al-go-rithms/blob/master/deep_learning/python/transfer-learning/image/transfer-learning.jpeg)
+# Transfer Learning
+In Transfer Learning, the knowledge of an already trained Machine Learning model is applied to a different but related problem. For
+example, if you trained a simple classifier to predict whether an image contains a backpack, you could use the knowledge that the 
+model gained during its training to recognize other objects like sunglasses. With transfer learning, we basically try to exploit 
+what has been learned in one task to improve generalization in another. We transfer the weights that a Network has learned at Task 
+A to a new Task B.<br>
+The general idea is to use knowledge, that a model has learned from a task where a lot of labeled training data is available, in a 
+new task where we don’t have a lot of data. Instead of starting the learning process from scratch, you start from patterns that 
+have been learned from solving a related task. Transfer Learning is mostly used in Computer Vision and Natural Language Processing 
+Tasks like Sentiment Analysis, because of the huge amount of computational power that is needed for them.
+![alt workflow](https://github.com/BAJUKA/al-go-rithms/blob/master/deep_learning/python/transfer-learning/image/transfer-learning.jpeg)

sentiment_analysis_twitter/Deep Learning/sentiment_cnn.py → machine_learning/sentiment_analysis_twitter/Deep Learning/sentiment_cnn.py

@@ -1,50 +1,50 @@
-import numpy as np
-import pandas as pd
-from keras.layers import Input, Dense, Bidirectional, Embedding, Dropout, Flatten
-from keras.layers import concatenate, SpatialDropout1D, GlobalAveragePooling1D, GlobalMaxPooling1D
-from keras.layers.convolutional import Conv1D
-from keras.layers.convolutional import MaxPooling1D
-from keras.models import Model
-from sklearn.model_selection import train_test_split
-from utils import *
-
-
-maxlen = 150
-max_features = 2500
-
-
-gop = pd.read_csv('Data/gop.csv')
-data = gop[['text','sentiment']]
-
-# Balance Negative - Positive tweets
-data[data['sentiment'] == 'Negative'] = data[data['sentiment'] == 'Negative'][:2236]
-data = data.dropna()
-
-data['sentiment'].value_counts() #Negative: 8493; Neutral: 3142; Positive: 2236
-X, Y = format_data(data, max_features, maxlen)
-X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.25, random_state=42)
-
-
-# Input shape
-inp = Input(shape=(maxlen,))
-
-# Embedding and CNN
-x = Embedding(max_features, 150)(inp)
-x = SpatialDropout1D(0.25)(x)
-x = Conv1D(filters=32, kernel_size=3, padding='same', activation='relu')(x)
-x = MaxPooling1D(pool_size=2)(x)
-x = Conv1D(filters=16, kernel_size=5, padding='same', activation='relu')(x)
-x = MaxPooling1D(pool_size=4)(x)
-x = Flatten()(x)
-
-# Output layer
-output = Dense(1, activation='sigmoid')(x)
-
-model = Model(inputs=inp, outputs=output)
-model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
-
-
-model.fit(X_train, Y_train, epochs=5, batch_size=32, verbose=1)
-
-results = model.predict(X_test, batch_size=1, verbose=1)
-run_test(results, Y_test)
+import numpy as np
+import pandas as pd
+from keras.layers import Input, Dense, Bidirectional, Embedding, Dropout, Flatten
+from keras.layers import concatenate, SpatialDropout1D, GlobalAveragePooling1D, GlobalMaxPooling1D
+from keras.layers.convolutional import Conv1D
+from keras.layers.convolutional import MaxPooling1D
+from keras.models import Model
+from sklearn.model_selection import train_test_split
+from utils import *
+
+
+maxlen = 150
+max_features = 2500
+
+
+gop = pd.read_csv('Data/gop.csv')
+data = gop[['text','sentiment']]
+
+# Balance Negative - Positive tweets
+data[data['sentiment'] == 'Negative'] = data[data['sentiment'] == 'Negative'][:2236]
+data = data.dropna()
+
+data['sentiment'].value_counts() #Negative: 8493; Neutral: 3142; Positive: 2236
+X, Y = format_data(data, max_features, maxlen)
+X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.25, random_state=42)
+
+
+# Input shape
+inp = Input(shape=(maxlen,))
+
+# Embedding and CNN
+x = Embedding(max_features, 150)(inp)
+x = SpatialDropout1D(0.25)(x)
+x = Conv1D(filters=32, kernel_size=3, padding='same', activation='relu')(x)
+x = MaxPooling1D(pool_size=2)(x)
+x = Conv1D(filters=16, kernel_size=5, padding='same', activation='relu')(x)
+x = MaxPooling1D(pool_size=4)(x)
+x = Flatten()(x)
+
+# Output layer
+output = Dense(1, activation='sigmoid')(x)
+
+model = Model(inputs=inp, outputs=output)
+model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
+
+
+model.fit(X_train, Y_train, epochs=5, batch_size=32, verbose=1)
+
+results = model.predict(X_test, batch_size=1, verbose=1)
+run_test(results, Y_test)

sentiment_analysis_twitter/Deep Learning/sentiment_utils.py → machine_learning/sentiment_analysis_twitter/Deep Learning/sentiment_utils.py

@@ -1,59 +1,59 @@
-import pandas as pd
-import numpy as np
-from keras.preprocessing.text import Tokenizer
-from keras.preprocessing.sequence import pad_sequences
-
-
-
-def format_data(data, max_features, maxlen):
-    data = data[data.sentiment != "Neutral"]
-    data = data.sample(frac=1).reset_index(drop=True)
-    data['text'] = data['text'].apply(lambda x: x.lower())
-
-    Y = to_numerical(data['sentiment'].values) # 0: Negative; 1: Positive
-    X = data['text']
-
-    remove_rt_url(X)
-
-    tokenizer = Tokenizer(num_words=max_features)
-    tokenizer.fit_on_texts(list(X))
-
-    X = tokenizer.texts_to_sequences(X)
-    X = pad_sequences(X, maxlen=maxlen)
-
-    return X, Y
-
-
-def to_numerical(d):
-    """Converts the categorical df[col] to numerical"""
-    _, d = np.unique(d, return_inverse=True)
-    return d
-
-
-def run_test(results, Y_validate):
-    pos_correct, neg_correct, total_correct = 0, 0, 0
-    _, (neg_count, pos_count) = np.unique(Y_validate, return_counts=True)
-
-    for i, r in enumerate(results):
-        if r > 0.5:
-            r = 1
-        else:
-            r = 0
-
-        if r == Y_validate[i]:
-            total_correct += 1
-            if r == 0:
-                neg_correct += 1
-            else:
-                pos_correct += 1
-
-
-    print('Positive Accuracy:', pos_correct/pos_count * 100, '%')
-    print('Negative Accuracy:', neg_correct/neg_count * 100, '%')
-    print('Total Accuracy:', total_correct/(pos_count + neg_count) * 100, '%')
-
-
-def remove_rt_url(df):
-    url = r'((https?):((//)|(\\\\))+([\w\d:#@%/;$()~_?\+-=\\\.&](#!)?)*)'
-    df.replace(regex=True, inplace=True, to_replace=r'^RT ', value=r'')
+import pandas as pd
+import numpy as np
+from keras.preprocessing.text import Tokenizer
+from keras.preprocessing.sequence import pad_sequences
+
+
+
+def format_data(data, max_features, maxlen):
+    data = data[data.sentiment != "Neutral"]
+    data = data.sample(frac=1).reset_index(drop=True)
+    data['text'] = data['text'].apply(lambda x: x.lower())
+
+    Y = to_numerical(data['sentiment'].values) # 0: Negative; 1: Positive
+    X = data['text']
+
+    remove_rt_url(X)
+
+    tokenizer = Tokenizer(num_words=max_features)
+    tokenizer.fit_on_texts(list(X))
+
+    X = tokenizer.texts_to_sequences(X)
+    X = pad_sequences(X, maxlen=maxlen)
+
+    return X, Y
+
+
+def to_numerical(d):
+    """Converts the categorical df[col] to numerical"""
+    _, d = np.unique(d, return_inverse=True)
+    return d
+
+
+def run_test(results, Y_validate):
+    pos_correct, neg_correct, total_correct = 0, 0, 0
+    _, (neg_count, pos_count) = np.unique(Y_validate, return_counts=True)
+
+    for i, r in enumerate(results):
+        if r > 0.5:
+            r = 1
+        else:
+            r = 0
+
+        if r == Y_validate[i]:
+            total_correct += 1
+            if r == 0:
+                neg_correct += 1
+            else:
+                pos_correct += 1
+
+
+    print('Positive Accuracy:', pos_correct/pos_count * 100, '%')
+    print('Negative Accuracy:', neg_correct/neg_count * 100, '%')
+    print('Total Accuracy:', total_correct/(pos_count + neg_count) * 100, '%')
+
+
+def remove_rt_url(df):
+    url = r'((https?):((//)|(\\\\))+([\w\d:#@%/;$()~_?\+-=\\\.&](#!)?)*)'
+    df.replace(regex=True, inplace=True, to_replace=r'^RT ', value=r'')
     df.replace(regex=True, inplace=True, to_replace=url, value=r'')