[Py3.5] Regression - Forecasting and Predicting

/*******************************************************************************************************************
-- Title : [Py3.5] Regression - Forecasting and Predicting
-- Reference : pythonprogramming.net
-- Key word : quandl sklearn scikit-learn linear_model linearregression csv datetimeindex fillna math.ceil
                  preprocessing.scale forecast predict ggplot matplotlib pyplot style 선형모델 선형 모델
                  선형회귀 선형 회귀 예측 추론 플롯 차트 cross-validation cross validation 
*******************************************************************************************************************/

■ Figures

■ Scripts

# -*- coding: utf-8 -*-
 
import quandl, math
import numpy as np
import pandas as pd
from sklearn import preprocessing, cross_validation, svm
from sklearn.linear_model import LinearRegression
import matplotlib.pyplot as plt
from matplotlib import style
import datetime
 
# ------------------------------
# -- Set Dataframe Option
# ------------------------------
pd.set_option('display.height', 1000)
pd.set_option('display.max_rows', 500)
pd.set_option('display.max_columns', 500)
pd.set_option('display.width', 1000)
 
 
# ------------------------------
# -- Read .CSV
# ------------------------------
# Scenario : 일자별 주식 자료를 통해 몇일 후 주가를 예측
# Ref : dbrang.tistory.com/1227
 
df = pd.read_csv("manipulated_wiki_google.csv")
 
print(df.head(5))
print("... read_.csv", "." * 100, "\n")
 
# -- 컬럼을 인덱스로 변경
df["Date"] = pd.DatetimeIndex(df["Date"])   # 인덱스를 DatetimeIndex로 만들기 위해 변경
df = df.set_index("Date")
 
print(df.head(5))
print("... change_col_to_idx", "." * 100, "\n")
 
 
# ------------------------------
# -- Create Label
# ------------------------------
# Ref : dbrang.tistory.com/1227
 
forecast_col = 'Adj. Close'
df.fillna(value=-99999, inplace=True)                 # replace NaN to -99999
forecast_out = int(math.ceil(0.01 * len(df)))         # len() : df row count, math.ceil : 가장 작은 정수 올림
 
print("len(df):", len(df))
print("0.01*len(df):", 0.01 * len(df))
print("math.ceil(0.01*len(df)):", math.ceil(0.01 * len(df)))
print("forecast_out:", forecast_out)
 
# -- Create Label
df['label'] = df[forecast_col].shift(-forecast_out)   # Label 생성위해 Row-Shift
 
print(df.head())
print(",,, add_label_&_shift_rows", "," * 100, "\n")
 
 
# ------------------------------
# -- Make Training and Testing Set
# ------------------------------
X = np.array(df.drop(['label'], 1))    # label만 빼고 Train Array로
X = preprocessing.scale(X)             # 기계학습 기능을 (-1)~(+1) 범위로 설정
X_lately = X [-forecast_out :]
X = X[:-forecast_out]
 
df.dropna(inplace=True)                # NaN rows 삭제
y = np.array(df['label'])              # Label을 Test Array로
 
# -- 모델 검증(cross-validation)
X_train, X_test, y_train, y_test = cross_validation.train_test_split(X, y, test_size=0.2)
 
print("X_train:", X_train[:5], "\n")
print("X_test:", X_test[:5], "\n")
print("y_train:", y_train[:20], "\n")
print("y_test:", y_test[:20])
print(";;; cross-validation", ";" * 100, "\n")
 
 
# ------------------------------
# -- Model Fitting
# ------------------------------
 
# -- Linear Regression(일반 처리)
clf = LinearRegression(n_jobs=-1)
clf.fit(X_train, y_train)
 
confidence = clf.score (X_test, y_test)
 
print(confidence)
print("^^^ linear_regression", "^" * 100, "\n")
 
 
# ------------------------------
# -- Forecasting & Predicting
# ------------------------------
forecast_set = clf.predict (X_lately)
 
print(forecast_set, confidence, forecast_out)
print("*** predictng", "*" * 100, "\n")
 
 
# ------------------------------
# -- Visualization
# ------------------------------
style.use('ggplot')
 
df['Forecast'] = np.nan  # NaN으로 필드 추가
 
last_date = df.iloc[-1].name
print("last_date:", last_date)
 
last_unix = last_date.timestamp()
print("last_unix:", last_unix)
 
one_day = 86400
next_unix = last_unix + one_day
print("next_unix:", next_unix)
 
for i in forecast_set:
    next_date = datetime.datetime.fromtimestamp(next_unix)
    next_unix += 86400
    df.loc[next_date] = [np.nan for _ in range(len(df.columns)-1)]+[i]
 
df['Adj. Close'].plot()
df['Forecast'].plot()
plt.legend(loc=4)
plt.xlabel('xlabel_Date')
plt.ylabel('ylabel_Price')
plt.show()

■ Files

manipulated_wiki_google.csv