퀀트 투자/Qlib

[Qlib] CH3. 플랫폼 구성-(1)Data

SKY-STONE 2024. 1. 9. 19:55

Microsoft Qlib Framework

Contents

  • Installation
  • Framework Overview
  • Main Component
    • Data
    • Train
    • Backtest
  • Examples

 

CH3. Main Components - (1)Data

Introduction
  • Qlib Framework의 구성들은 Quant Research에 용이하도록 제작하는 것이 제일 중요한 제작 의도임
  • 이에 Configuration 정의와 qrun 이라는 실행 명령어로만 쉽게 동작하게 UI에 많은 신경을 씀
  • 이번 장에서는 기본적인 Configuration 정의 방법과 구성 방법을 알아보고자 함
 

Workflow: Workflow Management — QLib 0.9.3.99 documentation

Before getting into details, here is a complete example of qrun, which defines the workflow in typical Quant research. Below is a typical config file of qrun. After saving the config into configuration.yaml, users could start the workflow and test their id

qlib.readthedocs.io

1. Inititialization Section

provider_uri: "~/.qlib/qlib_data/cn_data"
region: cn
  • provider_uri
    • The URI of the Qlib data. For example, it could be the location where the data loaded by get_data.py are stored.
  • region
    • If region == “us”, Qlib will be initialized in US-stock mode.
    • If region == “cn”, Qlib will be initialized in China-stock mode.
2. Data Section
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
    
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]
  • data_handler_config:  Data handler support user-customized processors to process data
  • dataset:  Dataset is responsible to prepare model-specific dataset from the processed data of Data Handler
3. Model Section
model:
    class: LGBModel
    module_path: qlib.contrib.model.gbdt
    kwargs:
        loss: mse
        colsample_bytree: 0.8879
        learning_rate: 0.0421
        subsample: 0.8789
        lambda_l1: 205.6999
        lambda_l2: 580.9768
        max_depth: 8
        num_leaves: 210
        num_threads: 20
  • model: model is designed to make the prediction score about stocks
4. Analysis Section
port_analysis_config: &port_analysis_config
    strategy:
        class: TopkDropoutStrategy
        module_path: qlib.contrib.strategy.strategy
        kwargs:
            topk: 50
            n_drop: 5
            signal: <PRED>
    backtest:
        limit_threshold: 0.095
        account: 100000000
        benchmark: *benchmark
        deal_price: close
        open_cost: 0.0005
        close_cost: 0.0015
        min_cost: 5
    record:
        - class: SignalRecord
          module_path: qlib.workflow.record_temp
          kwargs: {}
        - class: PortAnaRecord
          module_path: qlib.workflow.record_temp
          kwargs:
            config: *port_analysis_config
  • port_analysis_config
    • strategy: Portfolio Strategy is designed to adopt different portfolio strategies, which means that users can adopt different algorithms to generate investment portfolios based on the prediction scores of the Forecast Model.
    • backtest: After users specifying the models(forecasting signals) and strategies, running backtest will help users to check the performance of a custom model(forecasting signals)/strategy.
    • record: Qlib contains an experiment management system named QlibRecorder, which is designed to help users handle experiment and analyse results in an efficient way