Machine learning is used to recognise patterns in past events and make predictions about future events. For better understanding of the principle in this very simple example, a prediction is made about the success of a football team (Trask, 2019):


Input data includes:

  • Number of average toes per player smile
  • Number of games won so far in % (Win/Loss ratio)
  • Number of fans

Already trained artificial neural network with inputs

Already trained artificial neural network with multiple inputs


Weighted summation
Calculation of the weighted sum

  inputs     weights    local prediction
  (8.5   ·   0.1)   =       0.85          (toes)
  (0.65  ·   0.2)   =    +  0.13          (wlratio)
  (1.2   ·   0.0)   =    +  0.0           (fans)
final prediction    =       0.98

The inputs represent the information and weight represents the knowledge. It is clear, that predictions cannot be 100% correct.



                                                                
Reference list:

Trask, Andrew W. (2019): Grokking Deep Learning. Chapter 3. Introduction to neural pediction: forward propagation. Manning Publications Co.

Trask, Andrew W. (Sept 2018). Grokking-Deep-Learning: A Simple Neural Network Making a Prediction [Software]. Github. https://github.com/iamtrask/Grokking-Deep-Learning/blob/master/Chapter3%20-%20%20Forward%20Propagation%20-%20Intro%20to%20Neural%20Prediction.ipynb  (Accessed 5 May 2024)



Last modified: Wednesday, 15 May 2024, 12:17 PM