How to make an online summary ;)
Summazer is the free app able to "squeeze" a text or a web page and extract its juice, that is the sentences with the highest information content, to generate an automatic online summary of the processed text. The app incorporate a powerful spell checker that allows you to identify any incorrectly written words on the fly and, if necessary, correct them with a click on the suggested words.
In addition to highlighting the most important phrases, Summazer allows you to analyze, with a graph, the trend of semantic relevance within the text, in order to provide an immediate picture of the sections of the text to which you want to give greater evidence (typically a good text - especially if it is intended for the Web , should carry the salient parts at the beginning and at the end in order to immediately provide the reader with the meaning of the message you want to provide). Finally, Summazer reports an innovative indicator of the "summarizability" of the processed text (SoT).
If the summarizability (SoT) is low it may depend on the presence of spelling errors, excessive use of obsolete or foreign words, or the processed text is full of sentences that are not very related to each other from a semantic point of view (as happens, for example, if you summarize an entire web page, instead of selecting and pasting only the text that actually conveys the meaning of the article).
In addition to highlighting the most important phrases, Summazer allows you to analyze, with a graph, the trend of semantic relevance within the text, in order to provide an immediate picture of the sections of the text to which you want to give greater evidence (typically a good text - especially if it is intended for the Web , should carry the salient parts at the beginning and at the end in order to immediately provide the reader with the meaning of the message you want to provide). Finally, Summazer reports an innovative indicator of the "summarizability" of the processed text (SoT).
If the summarizability (SoT) is low it may depend on the presence of spelling errors, excessive use of obsolete or foreign words, or the processed text is full of sentences that are not very related to each other from a semantic point of view (as happens, for example, if you summarize an entire web page, instead of selecting and pasting only the text that actually conveys the meaning of the article).
The percentage depends on the length of the text to be summarised: for long texts it is advisable to set 30%, for very long texts to set 10%, for short texts 50% of the length of the text entered may be sufficient.
In order to optimise the quality of the summary to be processed, it is essential that the words in the text to be analysed are spelt correctly and recognised by the vocabulary of the selected language. By enabling the spell-checker, you will be able to identify any terms that are not spelled correctly or not recognised in the vocabulary of the selected language. If you believe that an unrecognised term should be added to Nelsenso.Net's vocabularies, you can add it. If this addition is approved, it will be available the next day.
Summazer returns the following outputs:
- It identifies the typology of the text elaborated, currently the recognized typologies include: narrative passages, argumentative texts and informative texts .
- Extracts the principal named entities (proper names, locations, etc ...) in the form of hashtags for which it is then possible to search for images using the "Unsplash" search engine ( Unsplash: Beautiful Free Images & Pictures) to enrich your post, article or essay.
- Highlight sentences with the highest information content in the elaborated text.
- Generates an abstract of the text based on the most relevant sentences.
- It also provides a series of graphs to highlight the "behind the scenes" of the application, that is: the analysis of the trend of the semantic relevance within the processed text and the statistical distribution (Gaussian) of the semantic relevance . Long and well-written texts will be characterized by narrow bell curves shifted far to the right with respect to the ordinate axis (i.e. texts with a high expected value of semantic relevance and few frequency values cluster around the average value). On the other hand, texts that cannot be summarized (because they have several paragraphs that are not semantically related to each other) are characterized by large bells centered with respect to the ordinate axis, or with a very low expected value.