Please use this identifier to cite or link to this item:
https://www.um.edu.mt/library/oar/handle/123456789/10981| Title: | Text synthesis of film reviews from user surveys |
| Authors: | Bonnici, Joseph |
| Keywords: | Text processing (Computer science) Motion pictures -- Reviews Surveys |
| Issue Date: | 2015 |
| Abstract: | This project consists in building a film review generator which uses data collected from a number of user surveys. The main goals are to build a survey generator which is able to dynamically generate a set of questions according to the needs and to also construct an engine which is able to generate an essay in English that describes various characteristics of a particular movie. The problem is worth solving because it explores a different way by which viewers can read a review before watching a movie that contains a collective opinion rather than the traditional way of reading a review submitted by a single critic. Data is collected from a survey which can be filled fast and easily. The survey questions were chosen after reading a number of reviews and extracting their key elements. Since people tend to answer only the first questions of a survey a priority function is used to put questions with the highest priority on top and questions with the least priority at the bottom. Priority is given to questions which are mostly needed for the review. This is determined by three parameters which are the weight, the count and the entropy. Each question had a weight bounded to it according to its importance. The other parameter, the count is the number of times the question has been answered. The entropy is the level of disorder in the answers. The level of disorder portrays the trend by which a question is being answered. After collecting enough data, the text generator kicks in and generates an essay. The process adopts the three main steps used by natural language generators. These three steps are: The Text Plan Construction, The Paragraph Builder and The Sentence Realizer. The Text Plan module constructs the skeleton of the essay according to the data of the survey. The Sentence Builder constructs the structure of the sentences by traversing a number of trees that contain different sentence structures. Finally, the Sentence Realizer uses rules of the English language grammar to convert a paragraph structure into a paragraph consisting of words of the English vocabulary. Various algorithms are used to generate the adequate adjectives, pronouns, synonyms etc. The essay builder also provides the possibility to generate a short or a long essay according to the specifications of the user. The result of all this is a system which given enough data, is able to generate a film review in English that reflects the opinion of various people that contributed to a survey. |
| Description: | B.SC.IT(HONS |
| URI: | https://www.um.edu.mt/library/oar//handle/123456789/10981 |
| Appears in Collections: | Dissertations - FacICT - 2015 Dissertations - FacICTAI - 2015 |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 15BSCIT029.pdf Restricted Access | 6.18 MB | Adobe PDF | View/Open Request a copy |
Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.
