Problem 3 PCA Using the Fish Market Data Consider again the data in the previous problem, focusing just on the Perch case. Remember that you only use the test data for the final evaluation. The model parameters must be found only with the training data. (a) Use multiple linear regression to find a linear model to predict the vertical length in cm of the Perch from the weight. Compare the performance of your approach with the solution to Problem 2. What do you observe? (b) Use the principles of PCA to reduce the dimensionality of the data to one dimension (the Perch data matrix without the length column). You might need to preprocess the data matrix to get the best result. You can check out the preprocessing steps in the Special Feature MLX file. In essence, you want to use the principles of PCA to find the appropriate X = XP (following the Special Feature MLX approach) where X here is your reduced Perch data and P is the matrix that you find from the eigenvectors of your (possibly preprocessed) data matrix. (c) (Use multiple linear regression to find a linear model to predict the vertical length in cm of the Perch from the compressed data contained in the that you compute from part (b). Compare the performance of your approach with the approach in (a) and the solution in Problem 2. What do you conclude?