python s1_deal_data.pypython s2_data_cut.pypython s3_feature_get.pypython s4_baseline.pypython s4_train_Linear.pypython s4_train_RF.pypython s4_train_MLP.pypython s4_train_RNN.pypython s4_train_LSTM.pypython s4_train_Emb_LSTM.pypython s5_get_indices.pypython s6_ablation_study.py
deal_data_feature/attr:属性变量版本特征deal_data_feature/onehot:one-hot 消融版本特征results:s4 的验证集预测结果args:最优参数models:模型文件indices:s5 生成的指标
默认会优先尝试以下地区*作物:
- 高州 * Lychee
- 徐闻 * Pineapple
- 化州 * Longan
- 丰顺 * Rice
- 大埔 * Tea
如果某些组合在特征过滤后不存在,会自动从候选列表补齐。