Development of a traffic analysis system based on neural networks

The purpose of this work: developing an intelligent system for analyzing and further predicting traffic
Relevance: as a result of the work, we get an intelligent system that can predict the number of connections according to the trained data.
Main tasks:
Creating and implementing a system based on neural networks
Conducting experiments
Analysis of the obtained data
The subject of research is the analysis of network traffic
The object of research is a neural network system for analyzing and predicting traffic

Neural network

Original data
Tuesday 23 September 2014 at 10 hours 11 minutes and 25 seconds several connections were recorded:

RESULTS

Conclusion
The work carried out includes a system that can predict the network load.
The intelligent system aims to assist in predicting network load.
The experiments of learning efficiency of this model, which showed that the best choice of method as a function of an error, a method called the method of least squares, namely the sum (for each node in the output layer) of the squared difference between the target and actual values of output nodes in the output layer.