Using the ideas from earlier in this part of the assignment, find theta_19_hat, which should be a numpy array of length 19 that minimizes the MSE. For this, your starting guess should be a vector of 19 zeros. If you start from a random vector, scipy minimize will probably get stuck. The mean squared error for your model should be less than 0.12. You will need to define a function mse_for_model_on_full_data(theta) similar to the mse_for_model_on_NBA_data(theta) function earlier in this part of the assignment. 80]: def mse_for_model_on_full_data(theta): return cel(df [ WOR'), predicted_probability_of_winning_gi ven_features (points_and_bias, theta) theta_19 hat ... mse_for_model_on_full_data(theta_19 hat) 181): grader. check("7a") (81): 0 of 1 tests passed Tests failed: ./tests/q7a. py Test code: >>> mae_for_model_on_full_data(theta_19 _hat) < 0.12 True Test result: Trying: me_for_model_on_full_data(theta_19_ hat) < 0.12 Expecting : True Line 1, in ./tests/q7a. py0 Failed example: