Determining semantic content of textual clusters
AssigneeServiceNow, Inc.
InventorBaskar Jayaraman
Life of this patent
Estimating the term…
Abstract
The embodiments herein provide a framework for and specific implementations of machine learning (ML) analysis of incident, online chat, knowledgebase, skills, and perhaps other types of databases. The ML techniques described herein may include various forms of semantic analysis of textual information in these databases, such as clustering, term frequency, word embedding, paragraph embedding, and potentially other techniques. Advantageously, use of ML in the specific ways described herein can provide insights into this textual information that otherwise would be impossible to determine in an accurate or concise fashion.
Claims20
A computational instance of a remote network management platform comprising:
- a database containing a plurality of incident reports, wherein the incident reports include text-based fields that document technology-related problems experienced by users of a managed network, and wherein a subset of the incident reports have been determined to be members of a cluster based on a similarity metric applied to at least some of the text-based fields;
- and
- one or more processors configured to:
- create a corpus of text from one or more of the text-based fields in the subset of the incident reports in the cluster;
- preprocess the corpus of text by removing stop-words and punctuation;
- map words in the corpus of text to word stems, wherein one or more of the words in the corpus of text is mapped to each word stem;
- determine a set of n word stems most frequently-appearing word stems in the corpus of text as mapped;
- determine a set of n words respectively corresponding to the set of n word stems, wherein each of the set of n words is a shortest word form of the words in the corpus of text that map to the respectively corresponding word stem;
- and
- provide a textual description for the cluster based on the set of n words.
A computer-implemented method comprising:
- creating a corpus of text from one or more text-based fields in a subset of incident reports, wherein a database stores a plurality of incident reports, wherein the incident reports include text-based fields that document technology-related problems experienced by users of a managed network, and wherein the subset of the incident reports have been determined to be members of a cluster based on a similarity metric applied to at least some of the text-based fields;
- preprocessing the corpus of text by removing stop-words and punctuation;
- mapping words in the corpus of text to word stems, wherein one or more of the words in the corpus of text is mapped to each word stem;
- determining a set of n word stems most frequently-appearing in the corpus of text as mapped;
- determining a set of n words respectively corresponding to the n word stems, wherein each of the set the n words is a shortest word form of the words in the corpus of text that map to the respectively corresponding word stem;
- and
- providing a textual description for the cluster based on the set of n words.
An article of manufacture including a non-transitory computer-readable medium, having stored thereon program instructions that, upon execution by a computing device, cause the computing device to perform operations comprising:
- creating a corpus of text from one or more text-based fields in a subset of incident reports, wherein a database stores a plurality of incident reports, wherein the incident reports include text-based fields that document technology-related problems experienced by users of a managed network, and wherein the subset of the incident reports have been determined to be members of a cluster based on a similarity metric applied to at least some of the text-based fields;
- preprocessing the corpus of text by removing stop-words and punctuation;
- mapping words in the corpus of text to word stems, wherein one or more of the words in the corpus of text is mapped to each word stem;
- determining a set of n word stems most frequently-appearing in the corpus of text as mapped;
- determining a set of n words respectively corresponding to the n word stems, wherein each of the set of n words is a shortest word form of the words in the corpus of text that map to the respectively corresponding word stem;
- and
- providing a textual description for the cluster based on the set of n words.
Dependencies are inferred from the claim text.