Feature selection is a process that selects a subset of features for model construction. Taking advantage of feature selection can improve model performance.
In this version, we provide several filter methods for feature
selection. Note that module works in a cascade manner where
selected result of filter A will be input into next filter B.
User should pay attention to the order of listing when
supplying multiple filters to filter_methods param in job configuration.
Below lists available input models and their corresponding filter methods with links to examples:
| Input Models | Filter Method |
|---|---|
| None | manual |
| Binning | iv_filter(threshold) iv_filter(top_k) iv_filter(top_percentile) |
| Statistic | statistic_filter |
Most of the filter methods above share the same set of configurable parameters. Below lists their acceptable parameter values.
| Filter Method | Parameter Name | metrics | filter_type | take_high |
|---|---|---|---|---|
| IV Filter | filter_param | "iv" | "threshold", "top_k", "top_percentile" | True |
| Statistic Filter | statistic_param | "max", "min", "mean", "median", "std", "var", "coefficient_of_variance", "skewness", "kurtosis", "missing_count", "missing_ratio", quantile(e.g."95%") | "threshold", "top_k", "top_percentile" | True/False |
-
iv_filter: Use iv as criterion to selection features.
- filter_type: Support three modes: threshold value, top-k and top-percentile.
- threshold value: Filter those columns whose iv is smaller than threshold. You can also set different threshold for each party.
- top-k: Sort features from larger iv to smaller and take top k features in the sorted result.
- top-percentile. Sort features from larger to smaller and take top percentile.
- select_federated: If set to True, the feature selection will be performed in a federated manner. The feature selection will be performed on the guest side, and the anonymously selected features will be sent to the host side. The host side will then filter the features based on the selected features from the guest side. This param is available in iv_filter only.
- threshold: The threshold value for feature selection.
- take_high: If set to True, the filter will select features with higher iv values. If set to False, the filter will select features with lower iv values.
- host_filter_type: The filter type for host features. It can be "threshold", "top_k", "top_percentile". This param is available in iv_filter only.
- host_threshold: The threshold value for feature selection on host features. This param is available in iv_filter only.
- host_top_k: The top k value for feature selection on host features. This param is available in iv_filter only.
- filter_type: Support three modes: threshold value, top-k and top-percentile.
-
statistic_filter: Use statistic values calculate from DataStatistic component. Support coefficient of variance, missing value, percentile value etc. You can pick the columns with higher statistic values or smaller values as you need.
- filter_type: Support three modes: threshold value, top-k and top-percentile.
- threshold value: Filter those columns whose statistic metric is smaller than threshold. You can also set different threshold for each party.
- top-k: Sort features from larger statistic metric to smaller and take top k features in the sorted result.
- top-percentile. Sort features from larger to smaller and take top percentile.
- threshold: The threshold value for feature selection.
- take_high: If set to True, the filter will select features with higher metric values. If set to False, the filter will select features with lower iv values.
- filter_type: Support three modes: threshold value, top-k and top-percentile.
-
manually: Indicate features that need to be filtered or kept.
- keep_col: The columns that need to be kept.
- filter_out_col: The columns that need to be dropped.
Besides, we support multi-host federated feature selection for iv filters. Starting in ver 2.0.0-beta, all data sets will obtain anonymous header during transformation from local file. Guest use iv filters' logic to judge whether a feature is left or not. Then guest sends result filter back to hosts. During this selection process, guest will not know the real name of host(s)' features.
