Reference no: EM133375718
Module - Collaborative Problem
The RBF NN algorithm is to be used for classification of the Iris data set. In this problem the following subproblems are to be completed:
Question 1. Using the RBF in Section 2 of the Neural Network documentation develop in pseudocode an RBF NN (no bias) system to accomplish the following steps:,
- Develop (pseudocode) an RBF NN train/training function.
- Develop (pseudocode) an RBF NN classi cation function.
Question 2. Optional no need to discuss collaboratively - Analyze your design,
- Calculate the running time of the system above in O-notation.,
- Calculate the total running time of the above system as T(n) with each line of pseudocode or code accounted for.,
- How does the total running time T(n) compare to the running time in O-notation?
Question 3. Implement your developed RBF NN using Python.,
- Train three two class models using the Iris data set as input training data, the Iris data will need to be reconfigured as a one-vs-all or one-vs-one data set.,
- Process the test data set to determine which class each test observation belongs to, in this problem you will simply use all 150 observations as your test data.,
- What is the classification accuracy of your design?
- If you had any misclassifications what was the cause of this, e.g, did the spread have an impact?"