機器學習範例策略

In [ ]:

!pip install finlab > log.txt
!pip install ta-lib-bin > log.txt
!pip install git+https://github.com/microsoft/qlib.git > log.txt

!pip install finlab > log.txt !pip install ta-lib-bin > log.txt !pip install git+https://github.com/microsoft/qlib.git > log.txt

In [1]:

from finlab import data
from finlab.ml import feature as mlf
features = mlf.combine({
  'pb': data.get('price_earning_ratio:股價淨值比'),
  'pe': data.get('price_earning_ratio:本益比')
}, resample='W')

features.head()

from finlab import data from finlab.ml import feature as mlf features = mlf.combine({ 'pb': data.get('price_earning_ratio:股價淨值比'), 'pe': data.get('price_earning_ratio:本益比') }, resample='W') features.head()

Out[1]:

		pb	pe
datetime	instrument
2010-01-10	1101	1.52	19.45
	1102	1.47	14.92
	1103	0.77	39.89
	1104	0.90	71.60
	1107	NaN	NaN

In [5]:

from finlab.ml import label as mll

label = mll.return_percentage(features.index, resample='W', period=12)

from finlab.ml import label as mll label = mll.return_percentage(features.index, resample='W', period=12)

In [6]:

import finlab.ml.qlib as q

is_train = features.index.get_level_values('datetime') < '2020-01-01'

X_train, y_train = features[is_train], label[is_train]
X_test = features[~is_train]

model = q.XGBModel()
model.fit(X_train, y_train)
y_pred = model.predict(X_test)

import finlab.ml.qlib as q is_train = features.index.get_level_values('datetime') < '2020-01-01' X_train, y_train = features[is_train], label[is_train] X_test = features[~is_train] model = q.XGBModel() model.fit(X_train, y_train) y_pred = model.predict(X_test)

[929:MainThread](2024-04-02 15:11:21,936) INFO - qlib.timer - [log.py:127] - Time cost: 0.554s | Loading data Done
INFO:qlib.timer:Time cost: 0.554s | Loading data Done
[929:MainThread](2024-04-02 15:11:22,010) INFO - qlib.timer - [log.py:127] - Time cost: 0.039s | DropnaLabel Done
INFO:qlib.timer:Time cost: 0.039s | DropnaLabel Done
[929:MainThread](2024-04-02 15:11:23,366) INFO - qlib.timer - [log.py:127] - Time cost: 1.345s | CSZScoreNorm Done
INFO:qlib.timer:Time cost: 1.345s | CSZScoreNorm Done
[929:MainThread](2024-04-02 15:11:23,372) INFO - qlib.timer - [log.py:127] - Time cost: 1.429s | fit & process data Done
INFO:qlib.timer:Time cost: 1.429s | fit & process data Done
[929:MainThread](2024-04-02 15:11:23,379) INFO - qlib.timer - [log.py:127] - Time cost: 1.997s | Init data Done
INFO:qlib.timer:Time cost: 1.997s | Init data Done
/usr/local/lib/python3.10/dist-packages/xgboost/core.py:160: UserWarning: [15:11:23] WARNING: /workspace/src/learner.cc:742: 
Parameters: { "n_estimators" } are not used.

  warnings.warn(smsg, UserWarning)

[0]	train-rmse:0.99957	valid-rmse:0.99967
[20]	train-rmse:0.99827	valid-rmse:0.99963
[40]	train-rmse:0.99762	valid-rmse:0.99988
[60]	train-rmse:0.99731	valid-rmse:1.00019

[929:MainThread](2024-04-02 15:11:30,023) INFO - qlib.timer - [log.py:127] - Time cost: 0.088s | Loading data Done
INFO:qlib.timer:Time cost: 0.088s | Loading data Done
[929:MainThread](2024-04-02 15:11:30,028) INFO - qlib.timer - [log.py:127] - Time cost: 0.000s | fit & process data Done
INFO:qlib.timer:Time cost: 0.000s | fit & process data Done
[929:MainThread](2024-04-02 15:11:30,033) INFO - qlib.timer - [log.py:127] - Time cost: 0.097s | Init data Done
INFO:qlib.timer:Time cost: 0.097s | Init data Done

In [12]:

from finlab.backtest import sim

close = data.get('price:收盤價')

position = y_pred[close > close.average(60)].is_largest(10)

sim(position, resample='Q')

from finlab.backtest import sim close = data.get('price:收盤價') position = y_pred[close > close.average(60)].is_largest(10) sim(position, resample='Q')

Out[12]:

<finlab.core.report.Report at 0x7d6181883340>