An insurance company is evaluating house insurance premiums for the next year in a given area. The company collects data on three features: Feature 1: Age - The age of the house in years. Feature 2: Size - The number of rooms in the house. Feature 3: Crime rate - This is an indicator for the average crime rate in the area. The company evaluates a sample of 800 houses in the area. Company's analyst uses linear programming to learn aggregation weights associated with a weighted arithmetic mean, OWA, a weighted power mean (p=0.5) and the Choquet integral. The analyst also calculate the arithmetic mean for comparison. The data were transformed to show that high score should be associated with high premium. All variables are transformed to [0,1] interval. The following results are obtained: Function AM WAM PM (p=0.5) OWA Learned weights Av. AE error 0.211, 0.412, 0.377 0.230, 0.482, 0.288 0.514.0.320.0.166 0.1112 0.108 0.1022 0.1078 RMSE Please make sure the answers needs to be in a sequential order separated by a comma, In case you are skipping (not answering) any question, please mention". (a) Which is the best fitting model? Please type "AM", "WAM", "PM" or "OWA". (b) Which variable is the most prominent in predicting the insurance premium? Please type "Feature 1", "Feature 2", or "Feature 3". (c) Do models tend towards high or low value? Please type "high value" or "low value" 0.1551 0.1478 0.144 0.1542