An analyst employed at an insurance company developed a k-NN classification model using data from their existing customers to predict which customers were likely to churn (i.e., move to a competitor) or stay loyal (i.e., continue as customer). To improve the performance of the model, the analyst further developed a RapidMiner process, utilising a relevant Loop operator to identify an optimal value of the hyper parameter k for the classifier. The analyst obtained the following results from RapidMiner for this optimisation process (see below).

iteration	k-NN.k	accuracy	kappa
0	63	0.705686	0.214989
1	87	0.67893	0.098436
2	57	0.70903	0.231135
3	8	0.67893	0.179604
4	23	0.688963	0.17811
5	49	0.688963	0.188188
6	19	0.692308	0.184417
7	98	0.688963	0.12372
8	57	0.70903	0.231135
9	95	0.695652	0.153765
10	7	0.688963	0.198022
11	70	0.705686	0.19481
12	25	0.695652	0.190762
13	41	0.695652	0.190762
14	20	0.685619	0.171843
15	4	0.682274	0.218634
16	15	0.692308	0.174132
17	9	0.675585	0.158431
18	28	0.692308	0.179307
19	29	0.698997	0.202147

1. With reference to the results and your knowledge of predictive model optimisation, identify the specific optimisation approach that the analyst followed. Explain your answer.

2 Discuss typical strengths of the optimisation approach identified in (a) above.

3 Discuss typical weaknesses of the optimisation approach identified in (a) above.

4 Suggest an appropriate alternative optimisation approach that analyst could follow. Explain your answer.

5 With reference to the results, what is the optimal value of the hyper parameter? Explain your answer.