Editing Final Report/Thesis 2015 (section)

==Conclusions==

The work undertaken throughout the project has fulfilled the key aims and objectives of the project including statistical analysis of likely language of origin of the Somerton Man code, the design and implementation of software to test the ''Rubaiyat of Omar Khayyam'' as a ''one-time pad'' in conjunction with a new ''key'' technique, and developed a search engine to discover possible n-grams contained within the code.  The group was successful in completing all tasks outlined in the proposal, with the exception of the proposed extension task to analyse the mass spectrometer data of the Somerton Man's hair.

Through this, the group was able to critically review and further the statistical analysis  of the likely language of origin of the Somerton Man code conducted by previous groups.  The 2015 group has improved upon the search for n-grams conducted by previous groups by increasing number of results and search speed.  In addition to this, the results collected through implementing the search engine are valuable for future groups to analyse for useful grams that could be linked to the Somerton Man code.  The group has also furthered the exploration from previous groups into the possibility that the ''Rubaiyat of Omar Khayyam'' was used as a ''one-time pad'' to encrypt the Somerton Man code by testing a new ''key'' technique.

The skills developed through undertaking this project include text processing and programming in a variety of languages including Java, MATLAB and Python.
The group has also thoroughly researched and learnt how to implement and evaluate statistical tehcniques including ''chi-squared testing'', ''p-value'' calculation and ''hypothesis testing'' and developed skills in using Microsoft Excel software to perform statistical analyses.

The main conclusions drawn from the project work include that the Somerton Man code was not created using the ''Rubaiyat of Omar Khayam'' as a ''one-time pad'' and the proposed method of using letter position within words as the ''key''.  Further analysis is required to obtain meaningful or useful combinations of grams from the results of the n-gram search.  The ''Universal Declaration of Human rights'' has too small a sample size of words in each language to accurately represent the initial letter frequency in each language for use in ''chi-squared testing''.  Finally, although the results from the ''hypothesis testing'' were somewhat inconclusive, the results of all of the ''chi-squared testing'' have lead to the conclusion that we can now say more confidently than ever that English was the most likely language from which the Somerton Man code was written, assuming it is an ''initialism''.

Despite being unable to decipher the Somerton Man code, the 2015 group has designed and implemented software that has furthered past work into the investigation and provided useful tools and resources to be utilised by future Honours students.

===Future Work===
The search engine in Task 2 contained a bug in the code that collected the number of years in which each gram occurred, rather than actual frequency of occurrence in each year.  This brings about a potential for reanalysis of the database, in particular the 5-gram data, using raw count through minor modification to the search engine code.

There is also the potential to run a limited search on the search engine results using a more sophisticated code than the one used by the 2015 group.  This could be used to generate more useful gram combinations to find commonly used English expressions or phrases that could be linked to the Somerton Man code.

According to previous group’s results and our results by using One-time pad method for task 3. Maybe One-time pad is not an appropriate method to decipher the Somerton Man code and Rubaiyat. Future group may use some new decryption methods instead of using One-time pad for task 3.  

Future students could extend the statistical analysis to perform ''hypothesis testing'' on all European languages in ''Project Gutenberg'', but an alternate method to the ''chi-squared testing'' performed would have to be utilised since the chi-squared values for the code against all other languages were too high to produce usable ''p-values''.

Another option would be to focus on English as the most likely language and statistically analyse the code against genres as conducted by the 2013 group <ref>L. Griffith and P. Varsos. (2013). Semester B Final Report 2013 – Cipher Cracking [online]. Available: https://www.eleceng.adelaide.edu.au/personal/dabbott/wiki/index.php/Semester_B_Final_Report_2013_-_Cipher_cracking</ref>, but using the chi-squared method as opposed to the squared difference and standard deviation methods adopted to consolidate or refute the conclusions drawn from their results .

Future groups could also extend the 2013 groups analysis of the mass spectrometer data collected using laser ablation of the Somerton Man's hair <ref>L. Griffith and P. Varsos. (2013). Semester B Final Report 2013 – Cipher Cracking [online]. Available: https://www.eleceng.adelaide.edu.au/personal/dabbott/wiki/index.php/Semester_B_Final_Report_2013_-_Cipher_cracking</ref>.  The hair was collected from the plaster bust of the Somerton Man made after his autopsy.  The concentration of isotopes found in the hair can be used to find out the environment the Somerton Man lived in in the lead up to his death.  Additional data from a separate hair has been taken since the 2013 group's analysis, as well as the concentration of isotopes in the plaster.  This could be used by a future group to crosscheck the data between separate hairs, as well as crosscheck the isotopes in the plaster versus the hair to see if the isotopes from the plaster may have diffused into the hair.