from sklearn import datasets

from sklearn.svm import SVC

from mlxtend.plotting import plot_decision_regions

import matplotlib.pyplot as plt

%matplotlib inline

​

# দুটো ফিচার নিচ্ছি

iris = datasets.load_iris()

X = iris.data[:, [0, 2]]

y = iris.target

​

# ক্লাসিফায়ার ট্রেনিং করছি

svm = SVC(C=0.5, kernel='linear')

svm.fit(X, y)

​

​

# প্লট করছি নতুন লাইব্রেরি দিয়ে

plot_decision_regions(X, y, clf=svm, legend=2)

​

# দু পাশের লেখাগুলো সেট করছি

plt.xlabel('sepal length [cm]')

plt.ylabel('petal length [cm]')

plt.title('SVM on Iris')

plt.show()

import pandas as pd

import numpy as np

import sklearn

​

from sklearn.model_selection import train_test_split

sklearn.__version__

'0.21.3'

try:

  # %tensorflow_version only exists in Colab.

  # শুধুমাত্র জুপিটার নোটবুক/কোলাবে চেষ্টা করবো টেন্সর-ফ্লো ২.০ এর জন্য

  %tensorflow_version 2.x

except Exception:

  pass

TensorFlow 2.x selected.

# df = pd.read_csv('geoloc_elev.csv')

df = pd.read_csv('https://raw.githubusercontent.com/raqueeb/TensorFlow2/master/datasets/geoloc_elev.csv')

​

# আমাদের দুটো ফিচার হলেই যথেষ্ট 

X = df[['lat', 'lon']].values

y = df['target'].values

df.plot(kind='scatter',

        x='lat',

        y='lon',

        c='target',

        cmap='bwr');

# এখানে সাপোর্ট ভেক্টর মেশিন ব্যবহার করছি

from sklearn.svm import SVC

​

clf = SVC(gamma='auto')

clf.fit(X, y)

SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,

    decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf',

    max_iter=-1, probability=False, random_state=None, shrinking=True,

    tol=0.001, verbose=False)

# প্লটিং এর কিছু লাইব্রেরি ব্যবহার করছি, plot_decision_regions লাইব্রেরিটা বেশ ভালো

import matplotlib.pyplot as plt

from mlxtend.plotting import plot_decision_regions

plot_decision_regions(X=X, y=y, clf=clf, legend=2)

plt.xlabel("x", size=5)

plt.ylabel("y", size=5)

plt.title('SVM Decision Region Boundary', size=6)

plt.show()

X_train, X_test, y_train, y_test = train_test_split(X, y, 

    test_size = 0.3, random_state=0)

from sklearn.linear_model import LogisticRegression

lr = LogisticRegression()

lr.fit(X_train, y_train)

/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.

  FutureWarning)

​

​

​

​

​

LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,

                   intercept_scaling=1, l1_ratio=None, max_iter=100,

                   multi_class='warn', n_jobs=None, penalty='l2',

                   random_state=None, solver='warn', tol=0.0001, verbose=0,

                   warm_start=False)

hticks = np.linspace(-2, 2, 101)

vticks = np.linspace(-2, 2, 101)

aa, bb = np.meshgrid(hticks, vticks)

ab = np.c_[aa.ravel(), bb.ravel()]

​

c = lr.predict(ab)

cc = c.reshape(aa.shape)

​

ax = df.plot(kind='scatter', c='target', x='lat', y='lon', cmap='bwr')

ax.contourf(aa, bb, cc, cmap='bwr', alpha=0.5)

import tensorflow as tf

model = tf.keras.models.Sequential([

    tf.keras.layers.Dense(1, input_dim=2, activation='tanh'),

    tf.keras.layers.Dense(1, activation='sigmoid')

])

model.compile(tf.keras.optimizers.SGD(lr=0.5), 'binary_crossentropy', metrics=['accuracy'])

result = model.fit(X_train, y_train, epochs=20, validation_split=0.1)

Train on 945 samples, validate on 105 samples

Epoch 1/20

WARNING:tensorflow:From /tensorflow-2.0.0-rc2/python3.6/tensorflow_core/python/ops/nn_impl.py:183: where (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.

Instructions for updating:

Use tf.where in 2.0, which has the same broadcast rule as np.where

945/945 [==============================] - 1s 1ms/sample - loss: 0.6526 - accuracy: 0.6444 - val_loss: 0.6558 - val_accuracy: 0.6381

Epoch 2/20

945/945 [==============================] - 0s 65us/sample - loss: 0.6438 - accuracy: 0.6582 - val_loss: 0.6580 - val_accuracy: 0.6381

Epoch 3/20

945/945 [==============================] - 0s 56us/sample - loss: 0.6445 - accuracy: 0.6582 - val_loss: 0.6543 - val_accuracy: 0.6381

Epoch 4/20

945/945 [==============================] - 0s 62us/sample - loss: 0.6439 - accuracy: 0.6582 - val_loss: 0.6555 - val_accuracy: 0.6381

Epoch 5/20

945/945 [==============================] - 0s 62us/sample - loss: 0.6441 - accuracy: 0.6582 - val_loss: 0.6562 - val_accuracy: 0.6381

Epoch 6/20

945/945 [==============================] - 0s 60us/sample - loss: 0.6433 - accuracy: 0.6582 - val_loss: 0.6561 - val_accuracy: 0.6381

Epoch 7/20

945/945 [==============================] - 0s 65us/sample - loss: 0.6439 - accuracy: 0.6582 - val_loss: 0.6541 - val_accuracy: 0.6381

Epoch 8/20

945/945 [==============================] - 0s 64us/sample - loss: 0.6432 - accuracy: 0.6582 - val_loss: 0.6563 - val_accuracy: 0.6381

Epoch 9/20

945/945 [==============================] - 0s 62us/sample - loss: 0.6422 - accuracy: 0.6582 - val_loss: 0.6611 - val_accuracy: 0.6381

Epoch 10/20

945/945 [==============================] - 0s 61us/sample - loss: 0.6441 - accuracy: 0.6582 - val_loss: 0.6567 - val_accuracy: 0.6381

Epoch 11/20

945/945 [==============================] - 0s 61us/sample - loss: 0.6433 - accuracy: 0.6582 - val_loss: 0.6556 - val_accuracy: 0.6381

Epoch 12/20

945/945 [==============================] - 0s 58us/sample - loss: 0.6442 - accuracy: 0.6582 - val_loss: 0.6535 - val_accuracy: 0.6381

Epoch 13/20

945/945 [==============================] - 0s 64us/sample - loss: 0.6414 - accuracy: 0.6582 - val_loss: 0.6556 - val_accuracy: 0.6381

Epoch 14/20

945/945 [==============================] - 0s 59us/sample - loss: 0.6410 - accuracy: 0.6582 - val_loss: 0.6503 - val_accuracy: 0.6381

Epoch 15/20

945/945 [==============================] - 0s 59us/sample - loss: 0.6381 - accuracy: 0.6582 - val_loss: 0.6433 - val_accuracy: 0.6381

Epoch 16/20

945/945 [==============================] - 0s 60us/sample - loss: 0.6272 - accuracy: 0.6582 - val_loss: 0.6274 - val_accuracy: 0.6381

Epoch 17/20

945/945 [==============================] - 0s 62us/sample - loss: 0.6058 - accuracy: 0.6582 - val_loss: 0.6006 - val_accuracy: 0.6381

Epoch 18/20

945/945 [==============================] - 0s 62us/sample - loss: 0.5839 - accuracy: 0.6497 - val_loss: 0.5836 - val_accuracy: 0.6381

Epoch 19/20

945/945 [==============================] - 0s 60us/sample - loss: 0.5693 - accuracy: 0.6265 - val_loss: 0.5644 - val_accuracy: 0.6381

Epoch 20/20

945/945 [==============================] - 0s 66us/sample - loss: 0.5547 - accuracy: 0.6180 - val_loss: 0.5631 - val_accuracy: 0.6381

pd.DataFrame(result.history).plot(ylim=(-0.05, 1.05))

hticks = np.linspace(-2, 2, 101)

vticks = np.linspace(-2, 2, 101)

aa, bb = np.meshgrid(hticks, vticks)

ab = np.c_[aa.ravel(), bb.ravel()]

​

c = model.predict(ab)

cc = c.reshape(aa.shape)

​

ax = df.plot(kind='scatter', c='target', x='lat', y='lon', cmap='bwr')

ax.contourf(aa, bb, cc, cmap='bwr', alpha=0.5)

model = tf.keras.models.Sequential([

    tf.keras.layers.Dense(4, input_dim=2, activation='tanh'),

    # এই লেয়ার পরে যোগ করে আমরা দেখবো 

    # tf.keras.layers.Dense(4, activation='tanh'),

    tf.keras.layers.Dense(1, activation='sigmoid')

])

model.compile(tf.keras.optimizers.SGD(lr=0.5), 'binary_crossentropy', metrics=['accuracy'])

h = model.fit(X_train, y_train, epochs=20, validation_split=0.1)

Train on 945 samples, validate on 105 samples

Epoch 1/20

945/945 [==============================] - 0s 445us/sample - loss: 0.6506 - accuracy: 0.6254 - val_loss: 0.6501 - val_accuracy: 0.6381

Epoch 2/20

945/945 [==============================] - 0s 66us/sample - loss: 0.6330 - accuracy: 0.6582 - val_loss: 0.6388 - val_accuracy: 0.6381

Epoch 3/20

945/945 [==============================] - 0s 61us/sample - loss: 0.5987 - accuracy: 0.6582 - val_loss: 0.5745 - val_accuracy: 0.8000

Epoch 4/20

945/945 [==============================] - 0s 65us/sample - loss: 0.5086 - accuracy: 0.7714 - val_loss: 0.4680 - val_accuracy: 0.9429

Epoch 5/20

945/945 [==============================] - 0s 65us/sample - loss: 0.3835 - accuracy: 0.9693 - val_loss: 0.3352 - val_accuracy: 0.9905

Epoch 6/20

945/945 [==============================] - 0s 63us/sample - loss: 0.2611 - accuracy: 0.9979 - val_loss: 0.2248 - val_accuracy: 1.0000

Epoch 7/20

945/945 [==============================] - 0s 60us/sample - loss: 0.1792 - accuracy: 1.0000 - val_loss: 0.1615 - val_accuracy: 1.0000

Epoch 8/20

945/945 [==============================] - 0s 62us/sample - loss: 0.1321 - accuracy: 1.0000 - val_loss: 0.1212 - val_accuracy: 1.0000

Epoch 9/20

945/945 [==============================] - 0s 61us/sample - loss: 0.1028 - accuracy: 1.0000 - val_loss: 0.0966 - val_accuracy: 1.0000

Epoch 10/20

945/945 [==============================] - 0s 68us/sample - loss: 0.0838 - accuracy: 1.0000 - val_loss: 0.0802 - val_accuracy: 1.0000

Epoch 11/20

945/945 [==============================] - 0s 61us/sample - loss: 0.0705 - accuracy: 1.0000 - val_loss: 0.0678 - val_accuracy: 1.0000

Epoch 12/20

945/945 [==============================] - 0s 61us/sample - loss: 0.0607 - accuracy: 1.0000 - val_loss: 0.0588 - val_accuracy: 1.0000

Epoch 13/20

945/945 [==============================] - 0s 72us/sample - loss: 0.0533 - accuracy: 1.0000 - val_loss: 0.0519 - val_accuracy: 1.0000

Epoch 14/20

945/945 [==============================] - 0s 61us/sample - loss: 0.0473 - accuracy: 1.0000 - val_loss: 0.0464 - val_accuracy: 1.0000

Epoch 15/20

945/945 [==============================] - 0s 59us/sample - loss: 0.0426 - accuracy: 1.0000 - val_loss: 0.0419 - val_accuracy: 1.0000

Epoch 16/20

945/945 [==============================] - 0s 60us/sample - loss: 0.0387 - accuracy: 1.0000 - val_loss: 0.0383 - val_accuracy: 1.0000

Epoch 17/20

945/945 [==============================] - 0s 74us/sample - loss: 0.0355 - accuracy: 1.0000 - val_loss: 0.0352 - val_accuracy: 1.0000

Epoch 18/20

945/945 [==============================] - 0s 61us/sample - loss: 0.0328 - accuracy: 1.0000 - val_loss: 0.0324 - val_accuracy: 1.0000

Epoch 19/20

945/945 [==============================] - 0s 56us/sample - loss: 0.0303 - accuracy: 1.0000 - val_loss: 0.0301 - val_accuracy: 1.0000

Epoch 20/20

945/945 [==============================] - 0s 61us/sample - loss: 0.0283 - accuracy: 1.0000 - val_loss: 0.0282 - val_accuracy: 1.0000

from sklearn.metrics import confusion_matrix, classification_report

​

y_pred = model.predict_classes(X_test)

cm = confusion_matrix(y_test, y_pred)

​

pd.DataFrame(cm,

             index=["Miss", "Hit"],

             columns=['pred_Miss', 'pred_Hit'])

train_score = model.evaluate(X_train, y_train, verbose=0)[1]

test_score = model.evaluate(X_test, y_test,  verbose=0)[1]

​

print("""Accuracy scores:

   Train:\t{:0.3}

   Test:\t{:0.3}""".format(train_score, test_score))

Accuracy scores:

   Train:    1.0

   Test:    1.0

hticks = np.linspace(-2, 2, 101)

vticks = np.linspace(-2, 2, 101)

aa, bb = np.meshgrid(hticks, vticks)

ab = np.c_[aa.ravel(), bb.ravel()]

​

c = model.predict(ab)

cc = c.reshape(aa.shape)

​

ax = df.plot(kind='scatter', c='target', x='lat', y='lon', cmap='bwr')

ax.contourf(aa, bb, cc, cmap='bwr', alpha=0.5)