yuyasugano

def SMA_TaLib(df):
    df1 = df.copy()
    df1["ma5"] = talib.SMA(df1['close'], timeperiod=5)
    df1["ma15"] = talib.SMA(df1['close'], timeperiod=15)
    df1["diff"] = df1.ma5 - df1.ma15
    df1["unixtime"] = [datetime.timestamp(t) for t in df1.index]

    # line and Moving Average
    fig = plt.figure(figsize=(15,5))
    ax = fig.add_subplot(1, 1, 1)

yuyasugano

# pct_change
f = lambda x: 1 if x>0.0001 else -1 if x<-0.0001 else 0 if -0.0001<=x<=0.0001 else np.nan
y = df.rename(columns={'close': 'y'}).loc[:, 'y'].pct_change(1).shift(-1).fillna(0)
X = df.copy()
y_ = pd.DataFrame(y.map(f), columns=['y'])
df_ = pd.concat([df, y_], axis=1)

# check the shape
print('----------------------------------------------------------------------------------------')
print('X shape: (%i,%i)' % X.shape)

yuyasugano

from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.neighbors import KNeighborsClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score, f1_score

X_train, X_test, y_train, y_test = train_test_split(X_, y_, test_size=0.33, random_state=42)
print('X_train shape: {}'.format(X_train.shape))

yuyasugano

open = pd.Series(df['open'])
high = pd.Series(df['high'])
low = pd.Series(df['low'])
close = pd.Series(df['close'])
volume = pd.Series(df['volume'])

# pct_change for new column
X['diff'] = y

# Exponential Moving Average

yuyasugano

_ipyw_jlab_nb_ext_conf    0.1.0                    py37_0
_libgcc_mutex             0.1                        main
alabaster                 0.7.12                   py37_0
anaconda                  2019.07                  py37_0
anaconda-client           1.7.2                    py37_0
anaconda-navigator        1.9.7                    py37_0
anaconda-project          0.8.3                      py_0
asn1crypto                0.24.0                   py37_0
astroid                   2.2.5                    py37_0
astropy                   3.2.1            py37h7b6447c_0

yuyasugano

_ipyw_jlab_nb_ext_conf    0.1.0                    py37_0
_libgcc_mutex             0.1                        main
alabaster                 0.7.12                   py37_0
anaconda                  2019.07                  py37_0
anaconda-client           1.7.2                    py37_0
anaconda-navigator        1.9.7                    py37_0
anaconda-project          0.8.3                      py_0
asn1crypto                0.24.0                   py37_0
astroid                   2.2.5                    py37_0
astropy                   3.2.1            py37h7b6447c_0

yuyasugano

# feature selection
open = pd.Series(df['open'])
high = pd.Series(df['high'])
low = pd.Series(df['low'])
close = pd.Series(df['close'])
volume = pd.Series(df['volume'])

# pct_change for new column
X['diff'] = y

yuyasugano

# Univariate Statistics
from sklearn.feature_selection import SelectPercentile
select = SelectPercentile(percentile=25)
select.fit(X_train_full, y_train_full.values.ravel())
X_train_selected = select.transform(X_train_full)
X_test_selected = select.transform(X_test_full)
mask = select.get_support()
print(mask)
plt.matshow(mask.reshape(1, -1), cmap='gray_r')
plt.xlabel("Technical Indexes")

yuyasugano

# Model-based Selection
from sklearn.feature_selection import SelectFromModel
select = SelectFromModel(RandomForestClassifier(n_estimators=100, random_state=42),
                        threshold="1.25*mean")
select.fit(X_train_full, y_train_full.values.ravel())
X_train_model = select.transform(X_train_full)
print(X_train_model.shape)
X_test_model = select.transform(X_test_full)
mask = select.get_support()
print(mask)

yuyasugano

# Recursive Feature Elimination
from sklearn.feature_selection import RFE
select = RFE(RandomForestClassifier(n_estimators=100, random_state=42),
             n_features_to_select=15)
select.fit(X_train_full, y_train_full.values.ravel())
X_train_rfe = select.transform(X_train_full)
X_test_rfe = select.transform(X_test_full)
mask = select.get_support()
print(mask)
plt.matshow(mask.reshape(1, -1), cmap='gray_r')