You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.

test.sh 2.1 kB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748
  1. #!/bin/bash
  2. #modules="data_processing timeseries_processing feature_analysis detection_algorithms reinforcement"
  3. #modules="data_processing timeseries_processing"
  4. modules="detection_algorithm"
  5. #test_scripts=$(ls primitive_tests | grep -v -f tested_file.txt)
  6. for module in $modules
  7. do
  8. test_scripts=$(ls $module | grep -v -f tested_file.txt)
  9. #test_scripts=$(ls $module)
  10. for file in $test_scripts
  11. do
  12. for f in $tested_file
  13. do
  14. echo $f
  15. done
  16. echo $file
  17. # Test pipeline building
  18. #python primitive_tests/$file > tmp.txt 2>>tmp.txt
  19. python $module/$file > tmp.txt 2>>tmp.txt
  20. error=$(cat tmp.txt | grep 'Error' | wc -l)
  21. echo "\t#Pipeline Building Errors:" $error
  22. if [ "$error" -gt "0" ]
  23. then
  24. cat tmp.txt
  25. #rm tmp.txt
  26. break
  27. fi
  28. # Test on KPI dataset
  29. #python3 -m d3m runtime fit-produce -p pipeline.yml -r datasets/anomaly/kpi/TRAIN/problem_TRAIN/problemDoc.json -i datasets/anomaly/kpi/TRAIN/dataset_TRAIN/datasetDoc.json -t datasets/anomaly/kpi/TEST/dataset_TEST/datasetDoc.json -o results.csv -O pipeline_run.yml
  30. #python3 -m d3m runtime fit-produce -p pipeline.yml -r datasets/anomaly/kpi/TRAIN/problem_TRAIN/problemDoc.json -i datasets/anomaly/kpi/TRAIN/dataset_TRAIN/datasetDoc.json -t datasets/anomaly/kpi/TEST/dataset_TEST/datasetDoc.json -o results.csv 2>>tmp.txt
  31. # Test on Yahoo dataset
  32. #python3 -m d3m runtime fit-produce -p pipeline.yml -r datasets/anomaly/yahoo_sub_5/TRAIN/problem_TRAIN/problemDoc.json -i datasets/anomaly/yahoo_sub_5/TRAIN/dataset_TRAIN/datasetDoc.json -t datasets/anomaly/yahoo_sub_5/TEST/dataset_TEST/datasetDoc.json -o results.csv -O pipeline_run.yml
  33. python3 -m d3m runtime fit-produce -p example_pipeline.json -r ../datasets/anomaly/yahoo_sub_5/TRAIN/problem_TRAIN/problemDoc.json -i ../datasets/anomaly/yahoo_sub_5/TRAIN/dataset_TRAIN/datasetDoc.json -t ../datasets/anomaly/yahoo_sub_5/TEST/dataset_TEST/datasetDoc.json -o results.csv 2> tmp.txt
  34. error=$(cat tmp.txt | grep 'Error' | wc -l)
  35. echo "\t#Pipeline Running Errors:" $error
  36. if [ "$error" -gt "0" ]
  37. then
  38. cat tmp.txt
  39. #rm tmp.txt
  40. break
  41. fi
  42. echo $file >> tested_file.txt
  43. done
  44. done

全栈的自动化机器学习系统,主要针对多变量时间序列数据的异常检测。TODS提供了详尽的用于构建基于机器学习的异常检测系统的模块,它们包括:数据处理(data processing),时间序列处理( time series processing),特征分析(feature analysis),检测算法(detection algorithms),和强化模块( reinforcement module)。这些模块所提供的功能包括常见的数据预处理、时间序列数据的平滑或变换,从时域或频域中抽取特征、多种多样的检测算