#!/usr/bin/env python
# coding: utf-8

# # 확률적 경사 하강법

# <table align="left">
#   <td>
#     <a target="_blank" href="https://colab.research.google.com/github/rickiepark/hg-mldl/blob/master/4-2.ipynb"><img src="https://www.tensorflow.org/images/colab_logo_32px.png" />구글 코랩에서 실행하기</a>
#   </td>
# </table>

# ## SGDClassifier

# In[1]:


import pandas as pd

fish = pd.read_csv('https://bit.ly/fish_csv_data')


# In[2]:


fish_input = fish[['Weight','Length','Diagonal','Height','Width']].to_numpy()
fish_target = fish['Species'].to_numpy()


# In[3]:


from sklearn.model_selection import train_test_split

train_input, test_input, train_target, test_target = train_test_split(
    fish_input, fish_target, random_state=42)


# In[4]:


from sklearn.preprocessing import StandardScaler

ss = StandardScaler()
ss.fit(train_input)
train_scaled = ss.transform(train_input)
test_scaled = ss.transform(test_input)


# In[5]:


from sklearn.linear_model import SGDClassifier


# In[6]:


sc = SGDClassifier(loss='log_loss', max_iter=10, random_state=42)
sc.fit(train_scaled, train_target)

print(sc.score(train_scaled, train_target))
print(sc.score(test_scaled, test_target))


# In[7]:


sc.partial_fit(train_scaled, train_target)

print(sc.score(train_scaled, train_target))
print(sc.score(test_scaled, test_target))


# ## 에포크와 과대/과소적합

# In[8]:


import numpy as np

sc = SGDClassifier(loss='log_loss', random_state=42)

train_score = []
test_score = []

classes = np.unique(train_target)


# In[9]:


for _ in range(0, 300):
    sc.partial_fit(train_scaled, train_target, classes=classes)

    train_score.append(sc.score(train_scaled, train_target))
    test_score.append(sc.score(test_scaled, test_target))


# In[10]:


import matplotlib.pyplot as plt

plt.plot(train_score)
plt.plot(test_score)
plt.xlabel('epoch')
plt.ylabel('accuracy')
plt.show()


# In[11]:


sc = SGDClassifier(loss='log_loss', max_iter=100, tol=None, random_state=42)
sc.fit(train_scaled, train_target)

print(sc.score(train_scaled, train_target))
print(sc.score(test_scaled, test_target))


# In[12]:


sc = SGDClassifier(loss='hinge', max_iter=100, tol=None, random_state=42)
sc.fit(train_scaled, train_target)

print(sc.score(train_scaled, train_target))
print(sc.score(test_scaled, test_target))