Add baselines

.latent-data/qlib

@@ -1 +1 @@
-Subproject commit 88b0871c12d0b139da489c53e02444606f6ca634
+Subproject commit aa552fdb2089cf5b4396a6b75191d2c13211b42d

configs/qlib/workflow_config_naive_v1_Alpha360.yaml

@@ -30,8 +30,8 @@ port_analysis_config: &port_analysis_config
         min_cost: 5
 task:
     model:
-        class: NAIVE
+        class: NAIVE_V1
-        module_path: trade_models.naive_model
+        module_path: trade_models.naive_v1_model
         kwargs:
             d_feat: 6
     dataset:

configs/qlib/workflow_config_naive_v2_Alpha360.yaml

@@ -0,0 +1,64 @@
+qlib_init:
+    provider_uri: "~/.qlib/qlib_data/cn_data"
+    region: cn
+market: &market all
+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: []
+    learn_processors: []
+    label: ["Ref($close, -2) / Ref($close, -1) - 1"]
+port_analysis_config: &port_analysis_config
+    strategy:
+        class: TopkDropoutStrategy
+        module_path: qlib.contrib.strategy.strategy
+        kwargs:
+            topk: 50
+            n_drop: 5
+    backtest:
+        verbose: False
+        limit_threshold: 0.095
+        account: 100000000
+        benchmark: *benchmark
+        deal_price: close
+        open_cost: 0.0005
+        close_cost: 0.0015
+        min_cost: 5
+task:
+    model:
+        class: NAIVE_V2
+        module_path: trade_models.naive_v2_model
+        kwargs:
+            d_feat: 6
+    dataset:
+        class: DatasetH
+        module_path: qlib.data.dataset
+        kwargs:
+            handler:
+                class: Alpha360
+                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: {}
+        - class: SignalMseRecord
+          module_path: qlib.contrib.workflow.record_temp
+          kwargs: {}
+        - 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

exps/trading/baselines.py

@@ -5,7 +5,8 @@
 # python exps/trading/baselines.py --alg GRU        #
 # python exps/trading/baselines.py --alg LSTM       #
 # python exps/trading/baselines.py --alg ALSTM      #
-# python exps/trading/baselines.py --alg NAIVE      #
+# python exps/trading/baselines.py --alg NAIVE-V1   #
+# python exps/trading/baselines.py --alg NAIVE-V2   #
 #                                                   #
 # python exps/trading/baselines.py --alg SFM        #
 # python exps/trading/baselines.py --alg XGBoost    #
@@ -53,7 +54,8 @@ def retrieve_configs():
     # DoubleEnsemble: A New Ensemble Method Based on Sample Reweighting and Feature Selection for Financial Data Analysis, https://arxiv.org/pdf/2010.01265.pdf
     alg2names["DoubleE"] = "workflow_config_doubleensemble_Alpha360.yaml"
     alg2names["TabNet"] = "workflow_config_TabNet_Alpha360.yaml"
-    alg2names["NAIVE"] = "workflow_config_naive_Alpha360.yaml"
+    alg2names["NAIVE-V1"] = "workflow_config_naive_v1_Alpha360.yaml"
+    alg2names["NAIVE-V2"] = "workflow_config_naive_v2_Alpha360.yaml"
 
     # find the yaml paths
     alg2paths = OrderedDict()

lib/trade_models/naive_v1_model.py

@@ -0,0 +1,88 @@
+##################################################
+# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2021 #
+##################################################
+from __future__ import division
+from __future__ import print_function
+
+import random
+import numpy as np
+import pandas as pd
+
+from qlib.log import get_module_logger
+
+from qlib.model.base import Model
+from qlib.data.dataset import DatasetH
+from qlib.data.dataset.handler import DataHandlerLP
+
+
+class NAIVE_V1(Model):
+    """NAIVE Version 1 Quant Model"""
+
+    def __init__(self, d_feat=6, seed=None, **kwargs):
+        # Set logger.
+        self.logger = get_module_logger("NAIVE")
+        self.logger.info("NAIVE 1st version: random noise ...")
+
+        # set hyper-parameters.
+        self.d_feat = d_feat
+        self.seed = seed
+
+        self.logger.info("NAIVE-V1 parameters setting: d_feat={:}, seed={:}".format(self.d_feat, self.seed))
+
+        if self.seed is not None:
+            random.seed(self.seed)
+            np.random.seed(self.seed)
+        self._mean = None
+        self._std = None
+        self.fitted = False
+
+    def process_data(self, features):
+        features = features.reshape(len(features), self.d_feat, -1)
+        features = features.transpose((0, 2, 1))
+        return features[:, :59, 0]
+
+    def mse(self, preds, labels):
+        masks = ~np.isnan(labels)
+        masked_preds = preds[masks]
+        masked_labels = labels[masks]
+        return np.square(masked_preds - masked_labels).mean()
+
+    def model(self, x):
+        num = len(x)
+        return np.random.normal(loc=self._mean, scale=self._std, size=num).astype(x.dtype)
+
+    def fit(self, dataset: DatasetH):
+        def _prepare_dataset(df_data):
+            features = df_data["feature"].values
+            features = self.process_data(features)
+            labels = df_data["label"].values.squeeze()
+            return dict(features=features, labels=labels)
+
+        df_train, df_valid, df_test = dataset.prepare(
+            ["train", "valid", "test"],
+            col_set=["feature", "label"],
+            data_key=DataHandlerLP.DK_L,
+        )
+        train_dataset, valid_dataset, test_dataset = (
+            _prepare_dataset(df_train),
+            _prepare_dataset(df_valid),
+            _prepare_dataset(df_test),
+        )
+        # df_train['feature']['CLOSE1'].values
+        # train_dataset['features'][:, -1]
+        masks = ~np.isnan(train_dataset["labels"])
+        self._mean, self._std = np.mean(train_dataset["labels"][masks]), np.std(train_dataset["labels"][masks])
+        train_mse_loss = self.mse(self.model(train_dataset["features"]), train_dataset["labels"])
+        valid_mse_loss = self.mse(self.model(valid_dataset["features"]), valid_dataset["labels"])
+        self.logger.info("Training MSE loss: {:}".format(train_mse_loss))
+        self.logger.info("Validation MSE loss: {:}".format(valid_mse_loss))
+        self.fitted = True
+
+    def predict(self, dataset):
+        if not self.fitted:
+            raise ValueError("The model is not fitted yet!")
+        x_test = dataset.prepare("test", col_set="feature")
+        index = x_test.index
+
+        preds = self.model(self.process_data(x_test.values))
+        return pd.Series(preds, index=index)

lib/trade_models/naive_v2_model.py

@@ -17,8 +17,8 @@ from qlib.data.dataset import DatasetH
 from qlib.data.dataset.handler import DataHandlerLP
 
 
-class NAIVE(Model):
+class NAIVE_V2(Model):
-    """NAIVE Quant Model"""
+    """NAIVE Version 2 Quant Model"""
 
     def __init__(self, d_feat=6, seed=None, **kwargs):
         # Set logger.
@@ -29,8 +29,7 @@ class NAIVE(Model):
         self.d_feat = d_feat
         self.seed = seed
 
-        self.logger.info(
+        self.logger.info("NAIVE parameters setting: d_feat={:}, seed={:}".format(self.d_feat, self.seed))
-            "NAIVE parameters setting: d_feat={:}, seed={:}".format(self.d_feat, self.seed))
 
         if self.seed is not None:
             random.seed(self.seed)
@@ -61,10 +60,7 @@ class NAIVE(Model):
                 results.append(0)
         return np.array(results, dtype=x.dtype)
 
-    def fit(
+    def fit(self, dataset: DatasetH):
-        self,
-        dataset: DatasetH
-    ):
         def _prepare_dataset(df_data):
             features = df_data["feature"].values
             features = self.process_data(features)
@@ -83,8 +79,8 @@ class NAIVE(Model):
         )
         # df_train['feature']['CLOSE1'].values
         # train_dataset['features'][:, -1]
-        train_mse_loss = self.mse(self.model(train_dataset['features']), train_dataset['labels'])
+        train_mse_loss = self.mse(self.model(train_dataset["features"]), train_dataset["labels"])
-        valid_mse_loss = self.mse(self.model(valid_dataset['features']), valid_dataset['labels'])
+        valid_mse_loss = self.mse(self.model(valid_dataset["features"]), valid_dataset["labels"])
         self.logger.info("Training MSE loss: {:}".format(train_mse_loss))
         self.logger.info("Validation MSE loss: {:}".format(valid_mse_loss))
         self.fitted = True

scripts/trade/baseline.sh

@@ -16,7 +16,7 @@ fi
 gpu=$1
 market=$2
 
-algorithms="NAIVE MLP GRU LSTM ALSTM XGBoost LightGBM SFM TabNet DoubleE"
+algorithms="NAIVE-V1 NAIVE-V2 MLP GRU LSTM ALSTM XGBoost LightGBM SFM TabNet DoubleE"
 
 for alg in ${algorithms}
 do