Editing Final Report/Thesis 2015 (section)

====Evaluation and Justification====
Unfortunately, the database used in the search engine was smaller in size than the preferred Microsoft option<ref>C. X. Zhai et al. (2010, July 19-23). Web N-gram Workshop [online]. Available: http://research.microsoft.com/en-us/events/webngram/sigir2010web_ngram_workshop_proceedings.pdf</ref>.  This could have provided a larger number of search results to be used for future analysis, but since the workflow utilised to access this database did not suit the application proposed by this project, the Microsoft option was deemed unusable.  Another limitation of using the Google N-Gram database was that it only includes grams that have appeared more than 40 times across the corpus<ref>Google Books. (2012 July). Ngram Viewer [Online]. Available: http://storage.googleapis.com/books/ngrams/books/datasetsv2.html</ref>.  There is room for improvement in this task by potentially fixing the bug in the search engine code to record frequency of occurrence of grams in each year to calculate the total frequency of each gram.  As mentioned, another improvement could be to optimise the gram combination code in order to obtain more useful or interesting results.  Overall, I believe Task 2 was successful in achieving its aims to crate a search engine that found regular expressions that could be linked to the Somerton Man code by using the initial letters of the code to find commonly used English phrases.  In addition to this, the search time relative to the number of results produced has clearly been largely improved when compared to the web crawler developed in previous years.