You are on page 1of 1

Sigmoid Model: This the third model that we used for the classification process.

Because the last two models have rigidity. They are also used for the classification
process but their modelwas straight line. That is basically used to classify easily linearly
separable data but sigmoid models can be used to classify a little flexible data. It uses a
sigmoid model that is flexible in nature. Basic six elements of the sigmoid model are
mentioned below.

Data: Sigmoid neurons can take real data as an input. It can take an integer as well as a
decimalvalue as an input. While MP neurons and perceptron can only take binary data as
an input. Aswell as its Output can also be any real number between 0 and 1.

We generally used a normalization method to convert the output into the range of 0 to 1.
By showing the output between 0 and 1 we can analyze how good or bad our output is.
That is not possible in other models. If it is closer to 0 or 1 that means our output will be
that much accurate and correct and if it is near 0.5 that means the chances of the wrong
output is more.

Task: It is also used to deal with the classification of data, based on the input. But as it
can take real data as an input it can also give output in the range of 0 to 1. It is basically a
provide a three dimensional classification of data that will provide more accuracy. We're
attempting to regress the input and output relationship in terms of likelihood. The output
of sigmoid is between 0 and 1, the sigmoid function can still be used for binary
classification problems by setting a threshold.

You might also like