International Journal of Science and Research (IJSR), India Online ISSN: 2319-7064
Volume 2 Issue 9, September 2013
www.ij...
International Journal of Science and Research (IJSR), India Online ISSN: 2319-7064
Volume 2 Issue 9, September 2013
www.ij...
International Journal of Science and Research (IJSR), India Online ISSN: 2319-7064
Volume 2 Issue 9, September 2013
www.ij...
of 3

Narrative Science: A Review

In this review work we developed a novel algorithm that can generate a narrative report of cricketing domain based on statistical data information extracted from raw values of match data. The algorithm is successful able to match quality standards of creating reporting with fair amount of increase in information gain.
Published on: Mar 3, 2016
Published in: Education      Technology      
Source: www.slideshare.net


Transcripts - Narrative Science: A Review

  • 1. International Journal of Science and Research (IJSR), India Online ISSN: 2319-7064 Volume 2 Issue 9, September 2013 www.ijsr.net Narrative Science: A Review Ramandeep Ghuman1 , Ripmi Kumari2 1, 2 Swami Vivekanand Institute of Engineering and Technology, Banur, Punjab, India Abstract: In this review work we developed a novel algorithm that can generate a narrative report of cricketing domain based on statistical data information extracted from raw values of match data. The algorithm is successful able to match quality standards of creating reporting with fair amount of increase in information gain. Keywords: Natural language processing, Narratives science, Data, Information gain. 1. Introduction Narrative science is branch of NLP which turns data into stories. Narrative is defined as some kind of retelling often in words (though it is possible to mime a story) of something that happened. Narrative meaning is about connections. It links individuals’ human actions and events into interrelated aspects of an understandable composite. Narrative science [1] is a kind of a tech solution that creates rich narrative content from data. Narrative is seamlessly created from structured data sources and can be fully customized to fit a customer’s voice, style and tone. stories are created in multiple formats, including long form stories ,headlines, tweets and industry reports with graphical visualizations. Multiple versions of the same story can be created to customize the content for each audience’s specific needs. In Narrative paradigm theory the all meaningful communication is a form of storytelling or giving a report of events and so human beings experience and comprehend life as a series of ongoing narratives, each with their own conflicts, character, beginnings, middles and ends. With help of narratives all forms of communication that appeal to our reason can best viewed as stories shaped by history culture and character and all forms of human communication are to be seen fundamentally as stories. The aim of narrative is to provide content and insight in those areas where it is either financially or logistically impossible for organizations to generate it themselves using traditional method. The figure below show how narrative science converts complex data into story.  In this first step, narrative science imports our data and builds an appropriate narrative structure to meet the goals of our audience.  In this second step create story using complex Artificial Intelligence algorithms..It extracts and organized key facts and insights and transformed them into stories into stories at scale.  In this step deliver insight, narrative science uses data to answer important questions, provide advice and deliver powerful insight in a precise clear narrative. 1.2 Advantages of Narrative Science Narrative provides the researcher with an understanding of data. Narrative gives the researcher access to stories or themes that the story teller may not even be conscious of Narrative highlights changing perspective and understanding of people and events as a function of time in the evaluation of an experience. Narrative science can help tackle a wide variety of business challenges and a broad range of company types. e.g Forbes.com uses narrative’s platform to generate corporate earnings preview stories. Narrative science helps companies leverage their data by automatically creating easy to use and consistent narrative reporting through our patented artificial intelligence platform. Narrative science turns big data into plain English and making sense of data. Big Data is defined as “data sets” whose size is beyond the ability of commonly used software tools to capture, manage and process the data within a tolerable elapsed time. Big Data is made of structured and unstructured information.10% structured information is the data in databases and is about 10% of story. Unstructured information is 90% of big data and is human information like emails, videos, tweets, facebook posts, call-center conversations, closed circuit television footage, mobile calls, website etc. Work with narrative science is easy to understand. With the help of narrative science we have more insights means deep knowledge in local language. Narrative science is more understanding and culture customized. 1.3 Applications or Why We Use Narrative Science We use narrative science to solve the problem of information explosion. People are essentially storytellers. Making decisions depends on judgments about ‘good reasons’. We are generating stories in the arenas of sports, finance, real estate and politics when people experience a story, the phase of comprehension is where people form a mental representation about the text. The mental representation that is formed is called a situation model while narrative science began its life providing content for Media Company Paper ID: 12013162 205
  • 2. International Journal of Science and Research (IJSR), India Online ISSN: 2319-7064 Volume 2 Issue 9, September 2013 www.ijsr.net 2. Methodology of Work Dataset of particular topic-In this we can take dataset of sports data like cricket match. The column of the table represents matches and each row corresponds to attribute like date of match, day, place, teams, bowling teams, batting teams etc of given matches. Each value is known as a datum. Extract statistical information-In this step we can extract statistical information or we can say measure data in the arithmetic mean which is an average value for a group of numerical observation. Develop formula base algorithm-From extract data or information we can develop formula based algorithm. This algorithm is rule base algorithm in which our said event recorded in dataset. Run story based on dataset-In this step narration story is given out as output..Story is based on the dataset. Algorithm extracts the key facts and interesting insights from the data and transforms them into stories. Calculate Information gain-Information gain is the expected reduction in entropy caused by partitioning the examples according to a given attributes. Entropy comes from information gain, higher thentropy the more the information content. Entropy is a common way to measure impurity. Entropy is very common in information theory characterizes the (im) purity of an arbitrary collection of examples. 3. Results Figure 1: Information gain Entropy defines the purity of Information of an arbitrary collection of examples. Information Gain is the expected reduction in entropy earned by partitioning the examples according to a given attribute. Figure 2: Information gain with narrative coefficient     Figure 3: Information gain value before adding narration and after adding narration Narrative Coefficient =Relevant Narrative Score⁄ Max. Narrative Score Figure 4 Information gain value before adding narration and after adding narration   Relavance Narration Score = Intro ɴ+Bodyɴ+Concɴ+Tittle ɴ Paper ID: 12013162 206
  • 3. International Journal of Science and Research (IJSR), India Online ISSN: 2319-7064 Volume 2 Issue 9, September 2013 www.ijsr.net Figure 5: Find the information gain ratio Information Gain Ratio = Information gain/Intrinsic value=IG/Iv 4. Conclusion All the results show that there is many folds of information gain when narration is added to dataset which are already based on the probability of events ,most of the reports, journals, articles, today’s seems to have ‘formula’ or ‘plot’ in their expression ,these ‘formula’ or ‘plots’ can also be understood in terms of their mathematical existence, especially when data set is analyzed on the basis of Descriptive statistics in which mean , average ,frequency, mode min and max values can be translated into meaningful ,along with its semantics narration which would be of machine quality but would be score fairly high in terms of qualitative perception of humans. By enhancing the information theory which is expressed in terms of probability now with narration of events captured in dataset, we were able to generate automatic report tool which generates the report in HTML format. References [1] Narrative Science, http://www.narrativescience.com (accessed in August 2012). [2] Lemke, J. L. (1990). Talking Science: Language, learning and Values. Norwood, NJ: Ablex Publishing. [3] Abell P (1987). The Syntax of Social Life: The Theory and Method of Comparative Narratives, New York, Oxford University Press. [4] Hempel CG (1965). Aspects of Scientific Explanation and Other Essaysin the Philosophy of Science, New York, Free Press. [5] Lowe NJ (2004). The Classical Plot and the Invention of Western Narrative, Cambridge University Press. [6] Polkinhorne DE (1988). Narrative knowing and the human sciences.New York: State University of New York Press. [7] Negrete A, Lartigue C (2004). Learning from Education to Communicate Science as a Good Story. Endevour. 28: 120-124. [8] PerinbanayagamR (1991). Discursive Acts. Hawthorne, NY: Aldine. Polkinhorne DE (1988). Narrative knowing and the human sciences. [9] Negrete A (2005). Facts via fiction stories that communicate science. In Sannit N (Ed.). Motivating Science.Science communication from a philosophical, educational and cultural perspective. Luthon pp. 95- 102. [10] Avraamidou, L.; Osborne, J. (2008) Science as Narrative: The story of the discovery of penicillin. The pantaneto forum home page. www.pantaneto.co.uk/issue31/ retrieved on 18september 2010. [11] Narrative Science, http://www.narrativescience.com (accessed in August 2012). [12] Nichols, N. Machine-Generated Content: Creating Compelling New Content from Existing Online. [13] Sources. PhD thesis, Northwestern University, June2010.GoogleAnalytics,http://www.google.com/ana lytics(accessed in August 2012) Author Profile Ramandeep Ghuman received the B. Tech degree in Computer Science Engineering from Punjabi university, Patiala in 2011 and Pursuing M. Tech degree in Computer Science Engineering from Swami Vivekanand Engineering Institute of Technology. Rimpi Kumara worked as Assistant Professor in Swami Vivekanand Institute of Engineering and Technology. She received her B. Tech degree in computer science from GNDU, Amritsar in 2009 and M. Tech degree in 2011 from GNDU, Amritsar. Paper ID: 12013162 207