From 09d751df598470bb4ca2c6ac2c9309f322781cde Mon Sep 17 00:00:00 2001
From: Caeden Lange <90806592+NotSkynet@users.noreply.github.com>
Date: Mon, 26 Jun 2023 18:28:54 -0700
Subject: [PATCH 1/2] Convert boolean features to float in training data

This commit addresses a type mismatch issue in the neural network training function. The one-hot encoded rank features in the training data were boolean values, which caused a UFuncTypeError when trying to perform operations with float values in the error term formula.

To fix this, we've added a line to convert all features in the training data to float. This converts the boolean rank features to numerical values (1.0 for True, 0.0 for False), which can be used in the error term formula without causing type errors.

We've also updated the markdown documentation to explain this conversion.
---
 .../student-admissions/StudentAdmissions.ipynb                 | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/intro-neural-networks/student-admissions/StudentAdmissions.ipynb b/intro-neural-networks/student-admissions/StudentAdmissions.ipynb
index 37b296a7d4..2b18669000 100644
--- a/intro-neural-networks/student-admissions/StudentAdmissions.ipynb
+++ b/intro-neural-networks/student-admissions/StudentAdmissions.ipynb
@@ -188,7 +188,7 @@
    "metadata": {},
    "source": [
     "## Splitting the data into features and targets (labels)\n",
-    "Now, as a final step before the training, we'll split the data into features (X) and targets (y)."
+    "Now, as a final step before the training, we'll split the data into features (X) and targets (y). We must also convert the values in the one hot encoded ranks from a boolean (true or false) to a float (1 or 0)."
    ]
   },
   {
@@ -201,6 +201,7 @@
     "targets = train_data['admit']\n",
     "features_test = test_data.drop('admit', axis=1)\n",
     "targets_test = test_data['admit']\n",
+    "features = features.astype(float)\n",
     "\n",
     "print(features[:10])\n",
     "print(targets[:10])"

From 93f00cb029d89034fe35025ac3b6fba186c7ec2f Mon Sep 17 00:00:00 2001
From: Caeden Lange <90806592+NotSkynet@users.noreply.github.com>
Date: Mon, 26 Jun 2023 18:54:53 -0700
Subject: [PATCH 2/2] changed the test accuracy aswell so it supports uses the
 astype(float)

---
 .../student-admissions/StudentAdmissions.ipynb                  | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/intro-neural-networks/student-admissions/StudentAdmissions.ipynb b/intro-neural-networks/student-admissions/StudentAdmissions.ipynb
index 2b18669000..73ba32d197 100644
--- a/intro-neural-networks/student-admissions/StudentAdmissions.ipynb
+++ b/intro-neural-networks/student-admissions/StudentAdmissions.ipynb
@@ -325,7 +325,7 @@
    "outputs": [],
    "source": [
     "# Calculate accuracy on test data\n",
-    "test_out = sigmoid(np.dot(features_test, weights))\n",
+    "test_out = sigmoid(np.dot(features_test.astype(float), weights))\n",
     "predictions = test_out > 0.5\n",
     "accuracy = np.mean(predictions == targets_test)\n",
     "print(\"Prediction accuracy: {:.3f}\".format(accuracy))"