Inkubate's ScoreIt! Tech: Identifying Writing Styles with NLP, Study notes of Technology

Download Inkubate's ScoreIt! Tech: Identifying Writing Styles with NLP and more Study notes Technology in PDF only on Docsity! THE AMAZING SCOREIT! TECHNOLOGY: The Story Behind the Science How Inkubate’s Algorithms Work to Identify Your Writing Style Inkubate’s technology is within the exciting and advanced field of “natural language processing.” When analyzing writing style, we are referring to the elemental building blocks and stylistic constructs of written communications. This is different and more meaningful than the concept of “writing style” that you may have studied in the classroom. For our purposes, your “writing style” is determined by analyzing and evaluating thousands of distinct patterns of writing choices that you have developed over the course of your lifetime. Your distinctive style of writing is a function of a number of factors including your gender, education, nationality, cultural influences, and personality traits. By analyzing your manuscript, or even a chapter or two of it, Inkubate’s ScoreIt!™ can very accurately map your writing style “fingerprint.” By applying this same science to all of the many authors in our corpus and, by comparing your writing style fingerprint to theirs, Inkubate’s ScoreIt!™ tool reports back to you the writers to whom you are best matched. ScoreIt!TM answers the question: “Who do I Write Like?” ScoreIt!™ aggregates thousands of proprietary writing style patterns into four major “families” to help you to more easily understand your personal writing style. These four families are as follows: Major Writing Style Feature Families Vocabulary The vocabulary and range used by the author to express ideas and the unique word combinations that describe the author’s literary “voice” Expressive Complexity The choice and distribution of word lengths. A large and complex vocabulary, for example, includes a variety of short and long words. Collectively, these word choices reflect an author’s distinctive expressive style. Grammar Describing complex human relationships typically requires more intricate grammar than describing a simple, action-based narrative. The range of word choices related to these “parts-of-speech” and the number and arrangement of prepositional phrases used reflect an author’s grammatical style. Tonal Quality This feature analyzes, among other aspects, the author’s use of “function words” such as “the”, “a/an” and “of”. Their use describes the functions of other words in a sentence and their frequency is a key characteristic of writing style. It is the overall mapping of these “features” and their intersections with those of the writers in our corpus, that allows us to evaluate and identify similar writing styles. Our corpus contains representative examples of virtually all the generally recognized fiction genre. By example, but by no means exhaustive, some of the major genre on this list include romance, fantasy, mystery, western, historical, and sci-fi. While Inkubate’s corpus is constantly evolving, most of the authors who published a bestselling book during the last 10 years, as well as many of the classic authors from the late 19th century and early part of the 20th century, are represented. Because ScoreIt!™ answers the intriguing question of “who do I write like?”, it is common that many of the authors who you write like will be writing within a genre close to yours. However, it is also common that some of the matched authors will be writing in other genres less closely aligned. That provides fascinating feedback and important insight that can help you identify who your audience may be and how you might target a wider audience through careful metadata selection. Genre classification and marketing keywords are examples of critically important metadata that enable your target audience to discover your book. Here, the analogy to the music service technologies previously mentioned is instructive. For instance, a person who loves the musical group Queen would likely be attracted to other musicians whom, at a macro level, could be easily classified into a variety of music genres including: rock, heavy metal, rock opera and perhaps others. But that person might also like classical music, jazz, opera or show tunes. Why is that? It is because music, like writing, is a form of communication. Your preferences are influenced by the sum total of your exposure to the vast variety of available communication content (in this case, music) and your unique and fluid reactions to that content. As you learn what you like, you learn to identify other new content that pleases you. While genre classifications are useful and help us to generally describe our likes and dislikes to each other, each piece of music is unique – just like each book. Your brain does not think about genre, it simply likes what it likes. It is limiting to think about your likes and dislikes by looking at genre classification alone. In the case of writing style, Inkubate leverages our ability to analyze why two unique writers are similar without considering the direct influence of genre, plot or “story arc.” This provide a new and valuable metric – writing style – that can be used to help writers find readers and readers to find writers. Furthermore, identifying writer similarities based upon these techniques has been proven by us to have important relevance and value in the marketplace. Consider the following well-known story: