diff --git a/3.evaluate_model/README.md b/3.evaluate_model/README.md
index 1829855d..c4594043 100644
--- a/3.evaluate_model/README.md
+++ b/3.evaluate_model/README.md
@@ -7,9 +7,12 @@ These predictions are saved in [model_predictions.tsv](evaluations/model_predict
We evaluate these 4 sets of predictions with a confusion matrix to see the true/false positives and negatives (see [sklearn.metrics.confusion_matrix](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.confusion_matrix.html) for more details).
-We also evaluate these 6 sets of predictions with [sklearn.metrics.f1_score](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html) to determine the final/shuffled baseline model's predictive performance on each subset.
+We also evaluate these 4 sets of predictions with [sklearn.metrics.f1_score](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html) to determine the final/shuffled baseline model's predictive performance on each subset.
F1 score measures the models precision and recall performance for each phenotypic class.
+In [class_PR_curves.ipynb](class_PR_curves.ipynb), we use [sklearn.metrics.precision_recall_curve](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_recall_curve.html) to derive the precision-recall curves for final and shuffled baseline models on the training and testing data subsets.
+The precision recall curves and their data are saved to [class_precision_recall_curves](class_precision_recall_curves/).
+
## Step 1: Evaluate Model
Use the commands below to evaluate the ML models:
diff --git a/3.evaluate_model/class_PR_curves.ipynb b/3.evaluate_model/class_PR_curves.ipynb
new file mode 100644
index 00000000..077b2e34
--- /dev/null
+++ b/3.evaluate_model/class_PR_curves.ipynb
@@ -0,0 +1,546 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Import Libraries"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import pandas as pd\n",
+ "import pathlib\n",
+ "from joblib import load\n",
+ "\n",
+ "import sys\n",
+ "sys.path.append(\"../utils\")\n",
+ "from split_utils import get_features_data\n",
+ "from train_utils import get_dataset\n",
+ "from evaluate_utils import class_PR_curves"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Load models and datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "model_dir = pathlib.Path(\"../2.train_model/models/\")\n",
+ "log_reg_model_path = pathlib.Path(f\"{model_dir}/log_reg_model.joblib\")\n",
+ "log_reg_model = load(log_reg_model_path)\n",
+ "shuffled_baseline_log_reg_model_path = pathlib.Path(f\"{model_dir}/shuffled_baseline_log_reg_model.joblib\")\n",
+ "shuffled_baseline_log_reg_model = load(shuffled_baseline_log_reg_model_path)\n",
+ "\n",
+ "# load features data from indexes and features dataframe\n",
+ "data_split_path = pathlib.Path(\"../1.split_data/indexes/data_split_indexes.tsv\")\n",
+ "data_split_indexes = pd.read_csv(data_split_path, sep=\"\\t\", index_col=0)\n",
+ "features_dataframe_path = pathlib.Path(\"../0.download_data/data/training_data.csv.gz\")\n",
+ "features_dataframe = get_features_data(features_dataframe_path)\n",
+ "\n",
+ "train_data = get_dataset(features_dataframe, data_split_indexes, \"train\")\n",
+ "test_data = get_dataset(features_dataframe, data_split_indexes, \"test\")\n",
+ "\n",
+ "save_dir = pathlib.Path(\"class_precision_recall_curves/\")\n",
+ "save_dir.mkdir(parents=True, exist_ok=True)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### PR Curves for combinations of models and data subsets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "(74256, 4)\n"
+ ]
+ },
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " Phenotypic_Class | \n",
+ " PR_Threshold | \n",
+ " Precision | \n",
+ " Recall | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " ADCCM | \n",
+ " 0 | \n",
+ " 0.033190 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " ADCCM | \n",
+ " 1 | \n",
+ " 0.033197 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " ADCCM | \n",
+ " 2 | \n",
+ " 0.033204 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " ADCCM | \n",
+ " 3 | \n",
+ " 0.033211 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " ADCCM | \n",
+ " 4 | \n",
+ " 0.033218 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "fig, PR_data = class_PR_curves(train_data, log_reg_model)\n",
+ "fig.suptitle(\"Precision Recall Curves for Final Model on Training Data\")\n",
+ "fig.savefig(f\"{save_dir}/final_model_train_data.png\")\n",
+ "\n",
+ "PR_data.to_csv(f\"{save_dir}/final_model_train_data.tsv\", sep=\"\\t\")\n",
+ "print(PR_data.shape)\n",
+ "PR_data.head()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "(74256, 4)\n"
+ ]
+ },
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " Phenotypic_Class | \n",
+ " PR_Threshold | \n",
+ " Precision | \n",
+ " Recall | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " ADCCM | \n",
+ " 0 | \n",
+ " 0.033190 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " ADCCM | \n",
+ " 1 | \n",
+ " 0.033197 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " ADCCM | \n",
+ " 2 | \n",
+ " 0.033204 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " ADCCM | \n",
+ " 3 | \n",
+ " 0.033211 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " ADCCM | \n",
+ " 4 | \n",
+ " 0.033218 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "fig, PR_data = class_PR_curves(train_data, shuffled_baseline_log_reg_model)\n",
+ "fig.suptitle(\"Precision Recall Curves for Shuffled Baseline Model on Training Data\")\n",
+ "fig.savefig(f\"{save_dir}/shuffled_baseline_model_train_data.png\")\n",
+ "\n",
+ "PR_data.to_csv(f\"{save_dir}/shuffled_baseline_model_train_data.tsv\", sep=\"\\t\")\n",
+ "print(PR_data.shape)\n",
+ "PR_data.head()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "(13136, 4)\n"
+ ]
+ },
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " Phenotypic_Class | \n",
+ " PR_Threshold | \n",
+ " Precision | \n",
+ " Recall | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " ADCCM | \n",
+ " 0 | \n",
+ " 0.032927 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " ADCCM | \n",
+ " 1 | \n",
+ " 0.032967 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " ADCCM | \n",
+ " 2 | \n",
+ " 0.033007 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " ADCCM | \n",
+ " 3 | \n",
+ " 0.033048 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " ADCCM | \n",
+ " 4 | \n",
+ " 0.033088 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "fig, PR_data = class_PR_curves(test_data, log_reg_model)\n",
+ "fig.suptitle(\"Precision Recall Curves for Final Model on Testing Data\")\n",
+ "fig.savefig(f\"{save_dir}/final_model_test_data.png\")\n",
+ "\n",
+ "PR_data.to_csv(f\"{save_dir}/final_model_test_data.tsv\", sep=\"\\t\")\n",
+ "print(PR_data.shape)\n",
+ "PR_data.head()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "(13136, 4)\n"
+ ]
+ },
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " Phenotypic_Class | \n",
+ " PR_Threshold | \n",
+ " Precision | \n",
+ " Recall | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " ADCCM | \n",
+ " 0 | \n",
+ " 0.032927 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " ADCCM | \n",
+ " 1 | \n",
+ " 0.032967 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " ADCCM | \n",
+ " 2 | \n",
+ " 0.033007 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " ADCCM | \n",
+ " 3 | \n",
+ " 0.033048 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " ADCCM | \n",
+ " 4 | \n",
+ " 0.033088 | \n",
+ " 1.0 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "