Editing Authorship detection: 2010 group (section)

===Semester 1, Week 5===
====Jie Dong====
'''Progress and Status this week:'''
# Algorithm update:
#* The new version of trigram extraction algorithm inserts a "#" before a sentence and a "$" after a sentence. For example, there is a string "Today is a good day. I want to go to picnic."After TextEditor class, it will becomes "# Today is a good day & # I want to go to picnic $"
#* The motivation to this modification is because in an English text, each sentence exists relatively independent with each other. In terms of the example above, "......a good day. I want ......", it is not necessary to calculate the probability of apperance of "I" after the bigram "good day". Instead, it will be more significant to characterise an author's writing habit by knowing the probability of apperance of "I" in the start of a sentence, i.e. after the bigram "$ #". Likewise, the probability of a word appearing at the end of the sentence is important to know as well, that is "day $ #".In addition, by this method, we can discover how often is a specific word used in one sentence
#* To determine the beginning and end of a sentence, delimiter "." is used. In the future, with further study of English text characteristics, there might be more delimiters
# Generate classification results based on Federalist Text.


'''Plan and Goals for new week:'''
# Perform more tests on different disputed texts
# Try another key words selection algorithm: based on occurring frequency

====Leng Tan====
'''Progress and Status this week:'''
Tried using Federalist Text.
best results give accuracy up to 70% when threshold = 10, data dimension = 25. this might be due to the short text length of the Federalist Text.
It is noted that WRI works better without normalization.
Found Greek File for the new Testament but not sure if is the right one.

'''Plan and Goals for new week:'''
Do Federalist Text again with different disputed text.
Try redo English text again with normalization.
====Tien-en Phua====
'''Progress and Status this week:'''
# Analysis of federalist result as it is most similar in style to the new testaments text 
# Namely that most of the federalist paper is written by Hamilton and likewise the new testaments is written by Paul with a few others written by different authors like Luke, John, Peter
# Comparison of results with other feature extraction algorithm
# After comparison of Function Word Analysis (FWA) and frequency occurrence of function words, the FWA proves to be a better algorithm as it produces more accurate results than frequency occurrence. 
# Using FWA reduces the need to chop text and allowing lesser data to be "chunk" out. 
'''Plan and Goals for new week:'''
# According to Gantt Chart, the implementation of controversies should take place next week. 
# Implement both FWA and frequency occurrence to the KJV text
# Frequency occurrence should produce consistent results to Talis.