The story plot, or narrative structure, forms the core of the story in classic novels or screenplays. On this basis, ideas and figures are combined to form a complex plot. It usually follows a very coarse structure in the form of a beginning, middle section and end. Usually, it is about a main character who is driven by a strong motivation (the dramatic question). In order to reach his or her goal, the character must overcome a number of obstacles placed in his or her path by an opponent (antagonist). There are numerous techniques and strategies to develop and redevelop the story plot.
In the case of natural language generation, the story plot also forms the structure of the text, but is much more detailed than that of traditional texts. And by its very nature, the story plot in the field of automated texts is also a less poetic affair – but technically, all the more complex. Here, data is processed in the form of places, names, rankings, statistics or contributions which were determined previously to be relevant. This means that they are organised, ordered and linked to certain conditions that make the text more complex.
Contrary to what some might expect from unrealistic ideas about artificial intelligence, the story plot is not created by an algorithm. A linguist develops it. For this purpose, he or she determines in advance the sequence of the individual data information in the form of prioritisations and makes an initial formatting in which paragraphs are defined at relevant points.
This ‘dramaturgy’ is then followed by the automatically generated text. On the basis of the structure, many different text variants can be produced that always contain new expressions of the same statement. This can be of great importance for SEO-relevant product descriptions, for example. Once the structure has been defined, the possibilities for variation are incredibly diverse.
https://www.dataversity.net/voice-recognition-and-natural-language-processing-how-do-they-work/