Summarization and Lexical Chaining

Busy professionals don’t have time to individually read through hundreds or thousands of long, tedious documents — you need summaries to quickly glance through and see what each document is about before going into deeper analysis.

Lexalytics has you covered: we’ll extract the most relevant sentences from each document so you can quickly understand the main ideas without spending valuable time reading the whole document. Lexalytics provides an abridged version of your content — 3 sentences by default, but customizable as desired — to supplement the analytic data from the other features of our text analysis software.

The summaries we provide are based on the words actually in the document. We give you the most important sentences. We can also give you the summaries that are relevant to an entity – great for dealing with 200 page analyst reports.

Lexalytics’ on-premise Salience Engine and in-the-cloud Semantria service summarize on a sentence-based level: we’ll select the sentences most pertinent to your contents’ topics, and then combine them to give you a concise synopsis of the original source text.

Like with theme extraction, Lexalytics uses lexical chaining to select the most relevant information to form a summary. Our software establishes a conceptual chain through related nouns, even when the concepts are from different areas of the document. The first sentence of each summary is the one that best represents the idea of your document, then continues with sentences ranked by decreasing length of their lexical association with the main topic.

We use the information pulled out of the text preparation stage (sentence breaking, lexical chaining, etc) and then combined with your parameters to provide the appropriate summary.

Lexalytics sifts out nonessential or unrelated sentences to show you the big ideas in a succinct manner. We provide summarization in all of our supported languages, and numerous customization options to best tailor your summaries to your individual preferences.