Run pytest test_module_18_challenge.py from the neural-network-challenge-1/ root directory
-
Two datasets were created: a target (y) dataset, which includes the "credit_ranking" column, and a features (X) dataset, which includes the other columns. (5 points)
-
The features and target sets have been split into training and testing datasets. (5 points)
-
Scikit-learn's StandardScaler was used to scale the features data. (5 points)
-
A deep neural network was created with appropriate parameters. (10 points)
-
The model was compiled and fit using the accuracy loss function, the adam optimizer, the accuracy evaluation metric, and a small number of epochs, such as 50 or 100. (10 points)
-
The model was evaluated using the test data to determine its loss and accuracy. (5 points)
-
The model was saved and exported to a keras file named student_loans.keras. (5 points)
-
The saved model was reloaded. (5 points)
-
The reloaded model was used to make binary predictions on the testing data. (10 points)
-
A classification report is generated for the predictions and the testing data. (10 points)
-
For Question 1:
-
The response describes the data that should be collected to build a recommendation system for student loan options. (4 points)
-
The response explains why they think that data should be collected. (4 points)
-
The type of data described is appropriate for a recommendation system for student loan options. (2 points)
-
-
For Question 2:
-
The response chose a filtering method. (4 points)
-
The student justified the choice of their filtering method. (4 points)
-
The choice of filtering method was appropriate for the data selected in the previous question. (2 points)
-
-
For Question 3:
-
The response lists two real-world challenges with building a recommendation system for student loans. (4 points)
-
The response explains why these challenges would be of concern for a student loan recommendation system. (6 points)
-