可以用少量股票的历史数据训练模型吗?（例如训练LGBModel）

❓ Questions and Help

我发现如果instruments中只填几只股票，例如随便选择的["SH600006", "SH600007"]

Training until validation scores don't improve for 50 rounds
[20] train's l2: 0.497512 valid's l2: 0.5
[40] train's l2: 0.497512 valid's l2: 0.5
Early stopping, best iteration is:
[1] train's l2: 0.497512 valid's l2: 0.5

然后用这个模型预测的话
pred_scores = model.predict(dataset)

datetime   instrument    score
2018-12-03  SH600006      0.0
            SH600007      0.0

我也使用了optuna 进行搜参，但是一直没有找到可用的参数组合

是我的使用方式有问题吗?

Start with the issue's two-instrument instruments example, LGBModel, and the model.predict(dataset) call. Check whether this training setup is supported and whether the reported early stopping and zero scores are expected; done means documenting the cause and the correct usage or configuration. The Optuna attempt is additional context.