if not os.path.exists(data_dir):
data_dir = '../data'
print(os.listdir(data_dir))
marcap_dir = os.path.join(data_dir, 'marcap')
marcap_data = os.path.join(marcap_dir, 'data')
os.listdir(marcap_data)
train_start = pd.to_datetime('2000-01-01')
train_end = pd.to_datetime('2020-06-30')
test_start = pd.to_datetime('2020-08-01')
test_end = pd.to_datetime('2020-10-31')
train_start, test_end
# 삼성전기 code '009150'
df_sem = read_marcap(train_start, test_end, ['009150'], marcap_data)
df_sem.drop(df_sem[df_sem['Marcap'] == 0].index, inplace=True)
df_sem.drop(df_sem[df_sem['Amount'] == 0].index, inplace=True)
df_sem['LogMarcap'] = np.log(df_sem['Marcap'])
df_sem['LogAmount'] = np.log(df_sem['Amount'])
df_sem
n_seq = 10
x_cols = ['LogMarcap', 'LogAmount', 'Open', 'High', 'Low', 'Close']
y_col = 'LogMarcap'
train_inputs, train_labels, test_inputs, test_labels, scaler_dic = load_datas_scaled(
df_sem, x_cols, y_col, train_start, train_end, test_start, test_end, n_seq)
train_inputs.shape, train_labels.shape, test_inputs.shape, test_labels.shape
model = build_model_rnn(n_seq, len(x_cols))
'028260', # 삼성물산
'032830', # 삼성생명
'034730', # SK
'035420', # NAVER
'051900', # LG생활건강
'051910', # LG화학
'055550', # 신한지주
'068270', # 셀트리온
'096770', # SK에너지
'105560', # KB금융
'207940', # 삼성바이오로직스
]
code_to_id = {code: i for i, code in enumerate(codes)}
# 삼성전기 code '009150'
df_sem = read_marcap(train_start, test_end, codes, marcap_data)
df_sem.drop(df_sem[df_sem['Marcap'] == 0].index, inplace=True)
df_sem.drop(df_sem[df_sem['Amount'] == 0].index, inplace=True)
df_sem.drop(df_sem[df_sem['Open'] == 0].index, inplace=True)
df_sem.drop(df_sem[df_sem['High'] == 0].index, inplace=True)
df_sem.drop(df_sem[df_sem['Low'] == 0].index, inplace=True)
df_sem.drop(df_sem[df_sem['Close'] == 0].index, inplace=True)
df_sem.drop(df_sem[df_sem['Volume'] == 0].index, inplace=True)
df_sem['LogMarcap'] = np.log(df_sem['Marcap'])
df_sem['LogAmount'] = np.log(df_sem['Amount'])
df_sem['LogOpen'] = np.log(df_sem['Open'])
df_sem['LogHigh'] = np.log(df_sem['High'])
df_sem['LogLow'] = np.log(df_sem['Low'])
df_sem['LogClose'] = np.log(df_sem['Close'])
df_sem['LogVolume'] = np.log(df_sem['Volume'])