# older_mortality **Repository Path**: Tomhappy/older_mortality ## Basic Information - **Project Name**: older_mortality - **Description**: No description available - **Primary Language**: Unknown - **License**: Not specified - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2023-12-10 - **Last Updated**: 2024-01-10 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # Step 1. Run all codes in the folder 'bigquery_data_extraction' # Please note that we used the MIMIC-III (v1.4), MIMIC-IV (v1.0), eICU-CRD (v1.2), Amsterdam-UMC (v0.1) # If you lack some basic codes, you need to look at the following links: # https://github.com/MIT-LCP/mimic-code # https://github.com/MIT-LCP/eicu-code # https://github.com/AmsterdamUMC/AmsterdamUMCdb 1. Please change 'db_name' to your project and database name 2. Run oasis_firstday_ams.sql -> sapsii_firstday_ams.sql -> older_score_ams.sql 3. Run sofa_eicu -> charlson_comorbidity_eicu.sql -> older_score_eicu.sql 4. Run older_score_mimiciv -> charlson_mimiciii.sql -> older_score_mimiciii.sql 5. Run sms_icu_score_all.sql # Step 2. Data proprocessing, Model development and evaluation # Please note that you need to change the paths where data and results are saved # # 1. Please change the setting: (1) older_model_main.py data_initial_path = 'D:/project/older_score/data/' result_path = 'D:/project/older_score/result/' (2) general_utils_data_processing.py Function: get_data_diff_source -> path (should the same as data_initial_path) Function: generate_xgb_gossis_sms -> data_basic_path (3) general_utils_model_score_eva.py Function: performance_scores -> apache_iv_info (4) general_utils_plot_table.py Function: basic_info_merge -> data_each (5) calibration_older_1/2.R result_path 2. Run older_model_main.py # calibration_older_1.R # calibration_older_2.R