RapidMiner and ANN skills may be required, but manipulation of data and software is not required. The results are in the attachment, please answer the questions according to the results.

(a) What is neural network analysis? What are the basic elements of neural networks? Why do companies use artificial neural network analysis? (20 marks) (b) From the performance vector of ANN output from RapidMiner calculate the accuracy and precision: (25 marks) Loyal Not Loyal Class precision Predicted Loyal 4802 1780 72.96% Predicted Not Loyal 628 3390 84.37% Class recall 88.43% 65.57%\fResults of Binary Logit Model or Logistic Regression Dependent Variable Encoding Original Value Internal Value Not loyal - O Loyal Model Summary -2 Log Step Likelihood Cox & Snell R Square Nagelkerke R Square 23.471 .453 .604 Estimation terminated at iteration number 6 because parameter estimates changed by less than .001. Classification Table Predicted Loyalty to the Brand Percentage Observed Not Loyal Loyal Correct Step 1 Loyalty to the Not loyal 12 brand 3 80.0 Loyal 3 12 80.0 Overall percentage 80.0 The cut Value is .500 Variables in the Equation B S. E. Wald df Step 1 Sig. Exp (B) Brand 1.274 479 7.075 - Product . 186 322 008 3.575 335 Shopping .590 .491 .563 1.205 1.442 - - - Constant -8.642 3.346 230 1.804 6.672 010 000 Variable(s) entered on step 1: Brand, Product, Shopping