Real-time neural network-based network analyzer for hotspot area

We present a real-time neural network-based network analyzer system for hotspot area. There are many applications that available in the market today for provide us the network graph of our hotspot areas. These graphs will be analyzed by a network administrator. Because a hotspot area often runs 24 hours, an administrator has a difficulty to monitor the traffic all the time. Therefore we proposed the automatic system to help a network administrator in monitoring the network. This system will replace human skill in interpreting the graph with an Artificial Neural Network System. To minimize the number of input vector we use mean value of axis, so the micro-computer e.g. notebook, laptop, PDA, and other gadgets can handle this system. Testing result showed this system could classify between normal, high and un-normal traffic of network graph periodically.