easypheno.postprocess.results_analysis
Module Contents
Functions
|
Summarize the results for each phenotype and datasplit for all models and save in a file. |
|
Convert result string saved in a .csv file to a dictionary |
|
Generate a heatmap based on the results summary .csv file |
- easypheno.postprocess.results_analysis.summarize_results_per_phenotype_and_datasplit(results_directory_genotype_level, eval_metric=None)
Summarize the results for each phenotype and datasplit for all models and save in a file.
The following files will be created:
at phenotype-folder level within results directories:
Detailed_results_summary_*PHENOTYPE*DATASPLIT-PATTERN*.xlsx: .xlsx-file containing detailed results for each phenotype and datasplit-maf pattern (e.g. with all runtime results etc.)
Results_summary**PHENOTYPE*DATASPLIT-PATTERN*.csv: .csv-file containing an overview of the performance of each model for a phenotype and datasplit-maf pattern combination
at genotype-folder level within results directories (the one that was specified):
Results_summary_all_phenotypes*DATASPLIT-PATTERN*.xlsx: .xlsx-file containing an overview of the performance of each model on each phenotype used for this genotype matrix with the specified datasplit-maf pattern
Results_summary_all_phenotypes*EVALMETRIC*DATASPLIT-PATTERN*.csv: only overview sheet of Results_summary*DATASPLIT-PATTERN*.xlsx
- Parameters
results_directory_genotype_level (str) – results directory at the level of the name of the genotype matrix
eval_metric (str) – eval metric to use for summary. Options (default given first): Regression: explained_variance, r2_score, rmse, mse. Classification: mcc, f1_score, accuracy, precision, recall.
- easypheno.postprocess.results_analysis.result_string_to_dictionary(result_string)
Convert result string saved in a .csv file to a dictionary