Qlib Double Ensemble for CSI 300 Stock Ranking
Summary
This Qlib configuration defines a Chinese equity research workflow using the CSI 300 universe and Alpha158 features. It trains a double ensemble of gradient-boosted models, with both sample reweighting and feature selection enabled, then evaluates signals through a top-k dropout portfolio strategy. The portfolio holds up to 50 stocks and replaces five positions as signals change.
The configuration separates training, validation, and test periods, with the portfolio backtest beginning in 2017 and using the Shanghai-listed CSI 300 index as its benchmark. It specifies closing-price execution, transaction costs, and a limit threshold. These settings make the experiment reproducible in outline, but the document contains no reported returns, comparison against alternatives, or robustness analysis. Results depend on the specified data, model settings, and backtest assumptions; the configuration alone does not establish that the strategy is profitable.
Key ideas
- The workflow trains a double ensemble using gradient-boosted trees and Alpha158 features.
- The model enables both sample reweighting and feature selection.
- A top-k dropout strategy ranks CSI 300 stocks and replaces a subset of holdings over time.
- The configuration specifies separate training, validation, and test periods and includes trading costs.
- No performance results or robustness checks are supplied.
Tags
Full text
# workflow_config_doubleensemble_Alpha158.yaml
```yaml
qlib_init:
provider_uri: "~/.qlib/qlib_data/cn_data"
region: cn
market: &market csi300
benchmark: &benchmark SH000300
data_handler_config: &data_handler_config
start_time: 2008-01-01
end_time: 2020-08-01
fit_start_time: 2008-01-01
fit_end_time: 2014-12-31
instruments: *market
port_analysis_config: &port_analysis_config
strategy:
class: TopkDropoutStrategy
module_path: qlib.contrib.strategy
kwargs:
signal: <PRED>
topk: 50
n_drop: 5
backtest:
start_time: 2017-01-01
end_time: 2020-08-01
account: 100000000
benchmark: *benchmark
exchange_kwargs:
limit_threshold: 0.095
deal_price: close
open_cost: 0.0005
close_cost: 0.0015
min_cost: 5
task:
model:
class: DEnsembleModel
module_path: qlib.contrib.model.double_ensemble
kwargs:
base_model: "gbm"
loss: mse
num_models: 3
enable_sr: True
enable_fs: True
alpha1: 1
alpha2: 1
bins_sr: 10
bins_fs: 5
decay: 0.5
sample_ratios:
- 0.8
- 0.7
- 0.6
- 0.5
- 0.4
sub_weights:
- 1
- 1
- 1
epochs: 28
colsample_bytree: 0.8879
learning_rate: 0.2
subsample: 0.8789
lambda_l1: 205.6999
lambda_l2: 580.9768
max_depth: 8
num_leaves: 210
num_threads: 20
verbosity: -1
dataset:
class: DatasetH
module_path: qlib.data.dataset
kwargs:
handler:
class: Alpha158
module_path: qlib.contrib.data.handler
kwargs: *data_handler_config
segments:
train: [2008-01-01, 2014-12-31]
valid: [2015-01-01, 2016-12-31]
test: [2017-01-01, 2020-08-01]
record:
- class: SignalRecord
module_path: qlib.workflow.record_temp
kwargs:
model: <MODEL>
dataset: <DATASET>
- class: SigAnaRecord
module_path: qlib.workflow.record_temp
kwargs:
ana_long_short: False
ann_scaler: 252
- class: PortAnaRecord
module_path: qlib.workflow.record_temp
kwargs:
config: *port_analysis_config
```Shown in full with attribution under the source's licence. Licence: MIT
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.