Shepperd's Pie

 This was bound to happen...Over-the-top gourmet SPie!

                                                   *     *     *

I've been on Machine Learnia #2 for a few days, getting myself back into

gangsta math. It's not easy to see how all this fits together. As W3schools

explains it, we are looking for an algorithm that will sort successfully for us!!

                                                   

*     *     *

The perceptron is a binary linear classifier.

The first step is to consider two features of plants, leaf width, lef height. We know

which ones are toxic; they have large and short leaves. Those that are naroow and long

are non-toxic. The perceptron we are putting together takes two inputs from each plant

\This is what we see on the graph.

                                        

Soon enough, we see there is a linear decision border.

Assigning weights to each factor, (and adding a bias), we can then tell whether

a new plant is toxic or not.

*     *     *

Next task, using an activation function to bring the outputs in the 0 to 1 range.

A commonly used one is the logistic function, in sigmoid presentation.

                                      

The original function dates from the 19th century, and describes a population

that grows exponentially until it exhausts its resources. It is related to the 

gaussian distribution of random outcomes, and covers all numbers...