AlphaForge(AFF)

Data Preparation

Similar to AlphaGen, We Use Qlib as data save tool and download data from free & open-source data source baostock.

Please install Qlib Qlib first

Then download stock data through running data_collection/fetch_baostock_data.py

The next, Modify the correspoding /path/for/qlib_data in gan.utils.data.py to the data you downloaded (the dafault setting is ~/.qlib/qlib_data/cn_data_rolling)

Run Our Model

stage1: Minning alpha factors

python train_AFF.py --instruments=csi300 --train_end_year=2020 --seeds=[0,1,2,3,4] --save_name=test --zoo_size=100

Here,

• instruments is the dataset to use, e.g., csi300,csi500.
• seeds is random seed list, e.g., [0,1,2] or [0].
• train_end_year is the last year of training set, when train_end_year is 2020,the train,valid and test set is seperately: 2010-01-01 to 2020-12-31,2021-01-01 to 2021-12-31,2022-01-01 to 2022-12-31
• save_name is the prefix when saving running results. zoo_size is the num of factors to save at stage 1 mining model.

stage2: Combining alpha factors

python combine_AFF.py --instruments=csi300 --train_end_year=2020 --seeds=[0,1,2,3,4] --save_name=test --n_factors=10 --window=inf

Here instruments,train_end_year,seeds,save_name,` must be the same as it in stage 1

• n_factors is the num of factors used at each day, it should be less than or equal to zoo_size in stage 1.
• window is the slicing window that is used to evaluate the alpha factors in order to dynamicly select and cobine.

stage3: Show the results

You could run the ipython notebook file

exp_AFF_calc_result.ipynb

to generate and concat experiment result.

Run baseline experiments

The experiment process of other models is similar to running our AFF model, Except that none of the other models have a combine step.

GP:

train: train_RL.py, show result: exp_RL_calc_result.ipynb

RL:

train: train_RL.py, show result: exp_RL_calc_result.ipynb

DSO:

train: train_RL.py, show result: exp_RL_calc_result.ipynb

ML models including XGBoost, LightGBM and MLP:

train & show results: exp_ML_train_and_result.ipynb