Pivot Point Based Intelligent System to Associate Creative Textual Artefacts

Readers many times look for similar books, similar movies or creative artefacts with reference to a particular creative artefact of their interest. This similarity is not simply about the subject but also about overall presentation and the way overall plot in a creative artefact evolves. Most of the systems generally consider meta-data, authors, titles and genres for recommending book(s) to readers or movie(s) to viewers. But they miss on a few crucial aspects of similarity including progression of sentiment or that of overall theme. Similarity is a holistic concept and not confined to the title or topic. In narratives, progression of emotions across the key events could prove to be the key for similarity. In this paper, we establish similarity among narratives based on computational relationships among sentiment progression using intelligent text analysis. For reference we have used database http://www.cs.cmu.edu/~dbamman/booksummaries.html with 16559 books. We have proposed Sentiment Progression Association Model (SP-AM) based on pivot points to derive and map similarities. Further, analysis of changes in sentiments is used to derive book clusters of interest to readers. The promising analytical findings endorse the scientific approach of deriving similarity based on sentiment progression. This pivot point based intelligent system can address numerous real-life problems related to recommendation, association and retrieval of creative artefacts.