Deep Learning(2)

1---Single input

Let's think about how single neurons work in data sets such as 80cereals.

Suppose a model:

Input：'sugars' （grams of sugars per serving）

Output：'calories' （calories per serving）

w：2.5

b：90

Looking back at the previous section(Deep Learning1), we can see the following images

In short, calories=sugars*2.5+90*1=5*2.5+90=102.5

2---Multiple inputs

80cereals isn't just 'sugars', it's something else, so how can we expand our model to include more things?

We can just add more input connections to the neuron.To find the output, we would multiply each input to its connection weight and then add them all together.

The formula for this neuron would be:

𝑦=𝑤0𝑥0+𝑤1𝑥1+𝑤2𝑥2+𝑏

A linear unit with two inputs will fit a plane, and a unit with more inputs than that will fit a hyperplane.