Qlib ALSTM Configuration for CSI 300 Alpha158 Signals
Summary
This configuration specifies a Qlib experiment that trains an attention-based LSTM model with a GRU recurrent layer on Alpha158 features for CSI 300 instruments. It selects 20 features, applies robust normalization and missing-value filling, and uses a forward return label comparing close prices across adjacent future sessions. The configured data spans 2008 through mid-2020, with training, validation, and test periods separated by date.
For portfolio analysis, the configuration uses a top-ranked portfolio with periodic dropouts and a CSI 300 benchmark. The backtest specifies close-price execution, transaction costs, a minimum fee, and a price-limit threshold. Signal, long-short analysis, and portfolio records are requested. These settings describe an experiment rather than its findings: the document includes no performance results, and its outcomes depend on the data, model implementation, and assumptions embedded in the configuration.
Key ideas
- The experiment trains an ALSTM model with a GRU layer on selected Alpha158 features for CSI 300 stocks.
- Feature processing includes robust z-score normalization and missing-value filling.
- The dataset separates training, validation, and test periods by date.
- Portfolio analysis uses a top-ranked strategy with a CSI 300 benchmark and specified trading costs.
- The configuration contains no results, so it does not show whether the model or strategy performed well.
Tags
Full text
# workflow_config_alstm_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
infer_processors:
- class: FilterCol
kwargs:
fields_group: feature
col_list: ["RESI5", "WVMA5", "RSQR5", "KLEN", "RSQR10", "CORR5", "CORD5", "CORR10",
"ROC60", "RESI10", "VSTD5", "RSQR60", "CORR60", "WVMA60", "STD5",
"RSQR20", "CORD60", "CORD10", "CORR20", "KLOW"
]
- class: RobustZScoreNorm
kwargs:
fields_group: feature
clip_outlier: true
- class: Fillna
kwargs:
fields_group: feature
learn_processors:
- class: DropnaLabel
- class: CSRankNorm
kwargs:
fields_group: label
label: ["Ref($close, -2) / Ref($close, -1) - 1"]
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: ALSTM
module_path: qlib.contrib.model.pytorch_alstm_ts
kwargs:
d_feat: 20
hidden_size: 64
num_layers: 2
dropout: 0.0
n_epochs: 200
lr: 1e-3
early_stop: 10
batch_size: 800
metric: loss
loss: mse
n_jobs: 20
GPU: 0
rnn_type: GRU
dataset:
class: TSDatasetH
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]
step_len: 20
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.