Editing Authorship detection: 2010 group (section)

===Semester 1, Week 6===
====Jie Dong====
'''Progress and Status this week:'''
# With the modification last week, I re-ran the test on English data set
# The classification accuracy increased to 85% - 90%. The highest was achieved when threshold = 30
# Clear trend can be observed: increasing size of training data, accuracy increases, threshold firstly increase and then drop
# Perform tests on Federalist paper, but the accuracy is very low, at about 35% average
# Discuss with supervisor and group member with the result on Federalist paper
# Since function words analysis achieve a good performance, it was suggested that combine part of them to enhance the algorithm
'''Plan and Goals for new week:'''
# Implement Trigram Markov model to select trigrams with "Golden Key words"
# Start to prepare form final seminar
# Achieve test results on King James Version
====Leng Tan====
'''Progress and Status this week:'''
# It was verified that the English text actually does not have any problem.
# Results shown was not favourable. The prediction was very inconsistent, achieving a low accuracy rate of 53% most of the time.
# It was suggested that the testing could be bias to Madison as only Madison text was taken as testing data.
# Comparing with the earlier results using the English text, which involves 100 training data and 70 disputed text, the accuracy and consistency was even much lower.
# WRI might not be suitable for authorship detection.
# It was suggested to combine the Function Word Frequency developed by Joel to enhance the algorithm.

'''Plan and Goals for new week:'''
# Examine the WRI algorithm with further testing.
# Implement enhanced version of the WRI by combining the algorithm with function word frequency.
# Prepare the powerpoint slides for the seminar.
# Start to make the initial stage of the video.

====Tien-en Phua====
'''Progress and Status this week:'''
# Analysis results of FWA on KJV
# Analysis results of frequency occurrence on KJV
# Frequency occurrence produces consistent results with Talis listing down Paul, Barnabas, Luke and Matthew as the possible authors
# FWA produces a different results which will be discuss WHY. 
# Discuss seminar structure with team
# Delegate task to team members for seminar 
# Produce a uniform set of data for testing and results presentation

'''Plan and Goals for new week:'''
# Consolidate results from English text, Federalist text and King James Version
# Research on future improvement for FWA
# Conduct a detail literature review on the background of the new testaments