Feature Selection Step#
To run locally: feature_selection
This feature selection step loads raw data that has been imputed, computes fitness metrics for all features in the raw data, and returns a subset of data (including list of fit features) for training.
Note: This step will be generalised and made available in a future release. Until it is done, this page is not updated.
The following is an example of a config file together with descriptions of its parts.
Step config example#
Alternatively, you can load an example of the filled config where you will need to change some values to adapt it to your problem.
1name: feature_selection
2
3input_data_specs:
4 datastore_type: influxdb
5 query_type: influx # influx/dataframe/stream/stream_dataframe/csv
6 query_template_path: ./configs/data/influx_query.txt
7 query_values:
8 start: "2020-05-20T13:30:00.000Z"
9 stop: "2020-05-20T13:35:00.000Z"
10 bucket: sensors-out
11 measurement: cat
12 tags: {}
13 data_converter:
14 name: influx_flat
15 args: {}
16
17output_data_specs:
18 - datastore_type: influxdb
19 settings:
20 bucket: test-bucket-1
21 measurement: testv1-fe
22
23
24run_specs:
25 input_sensors:
26 - "Load_Cell_Mid"
27 - "Eddy_Top"
28 data_filename: /Users/ngcs/Downloads/211106_220102_imputed.parquet
29 max_RUL: 1000
30 min_metric_val: 0.1
31 last_n_strokes: 10000
32 cycle_key:
33
Input and Output Data Specs#
input_data_specs
and output_data_specs
follow a standard format for all the pipeline
steps; see Octaipipe Steps.
Configs Description#
Use this table to describe in detail each of the fields in the configuration file provided above apart from input_data_specs and output_data_specs as they are explained in the main OctaiPipe Steps page.
Level 1 |
Level 2 |
Level 3 |
Type/Options |
Description |
---|---|---|---|---|
run_specs |
||||
Step Outputs#
In this section, provide a description of the expected outputs of the step, both local and in the blob to help users investigate their results.