Browse Prior Art Database

Optimal Solution to Identify the Recurring Incidents in Service Management Disclosure Number: IPCOM000244694D
Publication Date: 2016-Jan-06
Document File: 6 page(s) / 505K

Publishing Venue

The Prior Art Database


• For QA System, application of computation along with natural language processing would meet the cutting edge technologies, and thus help in business growth. • Instead of knowledge-based system, the proposed solution is based on ticket data information analyzed using standard methodologies of natural language processing. • There is no existing solution for identifying recurring incidents from natural language textual description, solution and other free textual filed of service management ticketing data / corpus, which is required to enhance the computation and automation for human machine interaction system.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 01 of 6

Optimal Solution to Identify the Recurring Incidents in Service Management

Background - Problem Statement

• Cumbersome activity of identifying recurring incidents.

• This activity is performed manually and utilizes a lot of time.

• This increases the probability of errors in the analysis as well.

Figure 1. Description and Steps

• The size of the data set adversely impact performance of the tool, however the tool should be scalable.

• The current set of tools are not ready to analyze unstructured data sets, often a large

portion of time is spent on cleaning and structuring the data as per individual account's need, which make data analysis more cumbersome.

• In order to structure a data set high effort and skills are required.

Figure 2. Description and Steps (Continued)


Page 02 of 6

Figure 3. Input and Output

 Upload / use data from ITSM ticketing tool for this analysis. Analyze CI and description field and Identify incidents occurring on the same CI. Filter down to the list of incidents occurring on a CI.

 Analyze free text description field to identify similar issue description. Narrow down to incidents occurring on same CI and similar descriptions using the following standard natural language processing techniques:



Page 03 of 6

o Preprocessing Data by Removing Stop Word

o Lemmatization

o Stemming

o Using Bag of Words

o Applying Term Frequency - Inverse Document Frequency

o Clustering Keywords

 Check if the Identified Issues have Same Restoration Steps by Analyzing Textual Solution Field.

 Check if the Shortlisted Issues have Same Cause.

Figure 4. Overall Flow Chart

Figure 5. NLP and ML Details


Page 04 of 6


The proposed method and apparatus is capable of

• Analyzing unstructured data like issue description, symptoms, solution, in free text or natural language and identify recurring incidents

• Extracting insights from unstructured data sources

• Automating the entire process

Comparison Study




• Analysis can be done for smaller amount of data
• Best suited for capacity reports / cause code analysis
• Pareto analysis is a method through which we can identify 80 % of issue through 20 % of criteria
• Pivot tables / Pivot charts can also highlight the focus areas for

problem management analysis

• Manual intervention...