ParkGunWook / ML_Flowering_Date

Using Azure Notebook_microsoft

ML_Flowering_Date

Using Azure Notebook_microsoft

Code Refer by https://www.tensorflow.org/tutorials/keras/basic_regression

1) Import Library For Machine Leaning

#Pandas library import for Excel Using
import pandas as pd 
#Numpy module import
import numpy as np 
import keras 
import matplotlib.pyplot as plt
import tensorflow as tf

from keras.models import Model, Sequential
from keras.layers import Dense

2) Data(Excel) Import with pandas, print and check Data(Excel)

Excelfile = pd.ExcelFile('dataset_for_Flowering_3.xlsx')
df = pd.read_excel(Excelfile)
df

3) Set Train data and test data and separate ‘result data’ to another space

Print and Check train and test Data(2-Dimensional array)

train_data = df.loc[1:104]
test_data = df.loc[105:112]

train_y = train_data['365percent']
train_data = train_data.drop('365percent', axis=1)

test_y = test_data['365percent']
test_data = test_data.drop('365percent', axis=1)

## 112 examples, 5 features
print("Training set: {}".format(train_data.shape))  
print("Testing set:  {}".format(test_data.shape))   

# Display sample features, notice the different scales
print(train_data.head(2)) 

4) Naming for data

column_names = ['Tem2', 'Tem3', 'Rain2', 'Rain3', 'Sun3']
df = pd.DataFrame(train_data, columns=column_names)
df.head(8)

5) Build Model(2-layer-64-32-1)

def build_model():
  model = keras.Sequential([
    keras.layers.Dense(64, activation=tf.nn.relu,
                       input_shape=(train_data.shape[1],)),
    keras.layers.Dense(32, activation=tf.nn.relu),
    keras.layers.Dense(1)
  ])

  optimizer = tf.train.RMSPropOptimizer(0.001)

  model.compile(loss='mse',
                optimizer= 'adam',
                metrics=['mae'])
  return model
  
model = build_model()
model.summary()

6) Display training progress with dot, Store data for progress working

# Display training progress by printing a single dot for each completed epoch
class PrintDot(keras.callbacks.Callback):
  def on_epoch_end(self, epoch, logs):
    if epoch % 100 == 0: print('')
    print('.', end='')

EPOCHS = 1000

# Store training stats
history = model.fit(train_data, train_y, epochs=EPOCHS,
                    validation_split=0.2, verbose=0,
                    callbacks=[PrintDot()])

7) Draw val loss with progress

def plot_history(history):
  f, ax = plt.subplots(figsize=(24,8))
  plt.xlabel('Epoch')
  plt.ylabel('Mean Abs Error')
  plt.plot(history.epoch, np.array(history.history['val_mean_absolute_error']),
           label = 'Val loss')
  plt.legend()
  plt.ylim([10, 30])

plot_history(history)

8) Draw and print test result

def plot_result():
  f, ax = plt.subplots(figsize=(24,8))
  plt.xlabel('Index')
  plt.ylabel('365Percent')
  plt.plot(test_data.index, model.predict(test_data),
           label='test_data predict')
  plt.plot(test_data.index, test_y,
           label = 'real test_data y')
  plt.legend()
  plt.ylim([70, 130])

plot_result()

print('365percent')
(test_data.index, model.predict(test_data))