Description
See [this document](../Practicum.md) for general information about the practicums.
Learning objectives:
– Naive Bayes classifier
Task 1: Implement a Naive Bayes classifier
– Load the Iris dataset and divide it into to 2/3 training and 1/3 test sets.
– Implement a Naive Bayes classifier
* a) Use categorical attributes by discretizing each attribute into three equally-sized bins: low, medium, high.
* b) Use continuous attributes and assume a Gaussian (normal) distribution. Estimate the parameters of the distribution (mean and variance) from the training data (you’ll have different parameters for each attribute)!
– Compare the performance of the two solutions in terms of accuracy and error rate. Fill in the results in the following table:
| Arrribute handling | Accuracy | Error rate |
| —————— | ——– | ———- |
| Discretization | | |
| Gaussian distr. | | |
References
– [Numpy arrays](http://docs.scipy.org/doc/numpy/reference/generated/numpy.array.html#numpy.array)
– [Numpy statistics](http://docs.scipy.org/doc/numpy/reference/routines.statistics.html)
