Abstract:
There are currently many type of industries that requires 24/7 monitoring on different levels for
smooth operations. Telecommunication is one of the industry where millions of alarms trigger on
daily basis from communication equipment’s and needs to be handled within time limit. These
monitoring operation is normally handled by human in loop which means huge amount of time
wastage. In order to minimize downtime, limit human control over this monitoring, companies
have implemented data mining and machine learning techniques that helps in not only proactive
monitoring of alarms along with suitable actions and also there is a huge time saved. In this
paper we have experimented with some real time telecommunication alarms that are gathered
from different telecommunication devices and occurred at different times. We have created a
system that can predict future occurrence of an alarm on the specified machine using machine
learning technologies. In this paper we have used decision tree classifier in order to classify huge
number of data received from devices. We are using it to predict alarms that are to be appeared
on a specific device/machine at an specific time stamp.
