Hello everyone.
In my team we are investigating how to build a new application that does "Root Cause Analysis" (similar to Machine Learning or Analytics) having as input the description and solution of tickets of an Application Maintenance Service.
The idea is to identify repetitive incidents (tickets) to launch improvement initiatives on these recurring incidents, before they occur.
We have the idea of identifying words or terms that are repeated in the fields of description of incidents. But we do not know how to do it.
We are novice developers with little experience in Splunk and we would appreciate all suggestions and advice on how to do this development, the existence of possible already developed app that we could use, or another Splunk solutions that are already developed and / or that we can improve or investigate .
Many thanks in advance for all the help and suggestions you can give us.
Q como va el proyecto?
@analiaeg - This is a great idea, but NLP (natural language processing) is a deep and wide subject. You are going to need a senior guy who understands the theory, or at least a mid-level guy who has been on this trip before. If your manager wants to do it cheaply, then he needs to give the lead to someone who can do it fast and right. Forty to sixty hours at $125 per hour is MUCH cheaper than 400-800 hours at $25/hour.
This general concept is known as Sentiment Analysis
and there are some apps on Splunkbase to give you a headstart:
https://splunkbase.splunk.com/apps/#/search/sentiment/
You will be able to implement this with Splunk. You need to ingest data from the Application Maintenance Service with one of many ways in which splunk can ingest data. Once data is ingested you can write search queries to create necessary dashboards.
Thanks for your answer.
Our problem is that we don't know what search queries we could write to identify repetitive incidents by identifying words or terms that are repeated in the description fields.
Could you give some ideas to investigate in this way?
Thank you very much.