超高频指数复现及策略回测 Framework Structure 已完成 | 领域 | 已完成 | status | [//]: # (|:-----------:|------------------------------------------------------------------------------------------------------|:------:|) | data | 计算合理freq | 1 | | data | 不要overlap | 1 | | data | optiver找因子，去极值，怎么划分区间（预测多少时间就按多少时间来切分切片），aggregate data | 1 | | data | 时间拉长（多日data），预测周期放大，0.2频率放大，多个股票一起训练，多日一起 | 1 | | model | 日内时间段切分，分别训练scaler/model | 1 | | model | 不同model-automl | 1 | | model | 防止过拟合-cv | 1 | | model | mljar适配 | -1 | | model | flaml适配 | 1 | | model | autogluon适配 | 1 | | model | rnn | 1 | | statistics | statistics胜率，盈亏比（每次盈亏） | 1 | | statistics | 因子IC,ICRank | 1 | | statistics | 多个股票上statistics平均表现 | 1 | | strategy | truncate小预测信号 | 1 | | backtest | 多只股票的通用模式 | 1 | | urgent | 1min跳的数据，每天240条数据 | 1 | | data | pca | -1 | | model | ensemble models, lasso, rnn, automl | 1 | | backtest | 方向正确率，\ 要排除连续到达阈值，且连续盈利的情况 | 1 | | backtest | 预测波动率为衍生品交易提供信息 | 1 | | data | 用lobster数据，复现 | -1 | | data | 去掉开收盘5min | 1 | | model | 过拟合1 Train Data Rows:530,Train Data Columns: 260 | 1 | | eda | y distribution, feature distribution

⚠ No verified equity curve — no track-record source connected.

No score history is stored yet — only the current score is shown.

⚠ No live verification account connected — ask for proof before buying.

Alerts on changes: coming soon

Prop-firm compatibility not provided.

Data-completeness & trust index (not a profitability rating)