October 1st - 5th

Week's Plan:
Continue working on the graph related work, continue making edits.
Work on polishing the proposal based on edits.
Get more reviewers.

Week's Accomplishments:
Added more content to graph section.

Problems:
Spent three days in the Children's Hospital with Wakako.

Next Week's Plan:

Continue working on the graph related work, continue making edits.
Work on polishing the proposal based on edits.
Get more reviewers.


Other Pieces of Work: (I just don't want to forget these ideas, not sure how necessary / important they are)

We should export the frequency data for the stoichiometry data and load that into yEd and see what see.

We should write some type of data loader that lets load in "hint-actions" so we can see where students request hints.

I should develop a Data-Properties class to more appropriately manage the different functions we have available depending on what type of data is read in. Basically the object just stores a dozen or so flags that are set based on what columns are read in, in the data-import stage of the program.

September 29th - October 2nd

Week's Plan:
Work on the related work section for graphs
Finish the first draft
Address the feedback received from Dr. Barnes.

Week's Accomplishments:
Made some edits.
Moved the document to google docs.
Worked on graph related work.

Problems:
Wakako was sick all week and so I had to stay home with her many days.

Next Week's Plan:
Continue working on the graph related work, continue making edits.

Other Pieces of Work: (I just don't want to forget these ideas, not sure how necessary / important they are)

We should export the frequency data for the stoichiometry data and load that into yEd and see what see.

We should write some type of data loader that lets load in "hint-actions" so we can see where students request hints.

I should develop a Data-Properties class to more appropriately manage the different functions we have available depending on what type of data is read in. Basically the object just stores a dozen or so flags that are set based on what columns are read in, in the data-import stage of the program.

September 17th - 21st

Week's Plan:
Add motivation to introduction section.
Read and write the rest of the sequence mining section.

Week's Accomplishments:
Added some motivation. Finished off the sequence mining section. Wrote up the interaction Network Section. 

Problems:
No problems, accomplished quite a bit.

Next Week's Plan:
Work on the graph section of the related work. Finish the first "rough draft" of the proposal.

Other Pieces of Work: (I just don't want to forget these ideas, not sure how necessary / important they are)

We should export the frequency data for the stoichiometry data and load that into yEd and see what see.

We should write some type of data loader that lets load in "hint-actions" so we can see where students request hints.

I should develop a Data-Properties class to more appropriately manage the different functions we have available depending on what type of data is read in. Basically the object just stores a dozen or so flags that are set based on what columns are read in, in the data-import stage of the program.

September 10th - 14th

Week's Plan:
Add motivation to introduction section.
Read and write the rest of the sequence mining section.

Week's Accomplishments:
Put my computer back together, so I can work from home again. Installed windows, etc.
Read some work on sequence mining.

Problems:
Had no internet.
Had no PC at home, but had parts come in.

Next Week's Plan:
Write up the interaction network section.
Finish the sequence mining related work section.
Add motivation to introduction.

Other Pieces of Work: (I just don't want to forget these ideas, not sure how necessary / important they are)

We should export the frequency data for the stoichiometry data and load that into yEd and see what see.

We should write some type of data loader that lets load in "hint-actions" so we can see where students request hints.

I should develop a Data-Properties class to more appropriately manage the different functions we have available depending on what type of data is read in. Basically the object just stores a dozen or so flags that are set based on what columns are read in, in the data-import stage of the program.

August 27th - 31st

Week's Plan:
Write a lot of my dissertation proposal. Focusing on combining like sections and filling out the proposed contributions section. Also add some more specific examples of the tables, diagrams, etc. so that it is more clear.

Week's Accomplishments:
I wrote a lot. I combined necessary sections and was able to make a concrete direction in the text. The proposed contributions chapter is more directed and concrete. The related work has the appropriate sections and direction, though lacks the text needed.

I submitted the work to Tiffany on Thursday night, for the Friday train ride. Also fixed our issues with dropbox.

Problems:
I was not able to address Entropy example. I worked on this but ran into two issues. For entropy  to work, we want to separate based on students that make it to a goal state and students that do not. If we use a simple data-set with less than 10 students, overlap is pretty rare, meaning most states have a frequency of 1. By default all frequency 1 states have an entropy of zero, meaning they are uninteresting. After you remove those states, the overall contribution of entropy is pretty low. However with a much larger graph, like 100 or 200 students, the number of states with a non-zero entropy should be significantly higher. 

I should graph the percent of non frequency one states, to case count (ie. sample size of students) to show the relationship of growth between these two variables. For large samples it seems roughly 50% of the states are frequency one. For smaller samples as much as 80 or 90 percent of states can be frequency one.

Next Week's Plan:
Read the feedback on my draft, make changes as necessary. Continue writing.
1) Add motivation to the introduction chapter.
2) Add content to related work.


Other Pieces of Work: (I just don't want to forget these ideas, not sure how necessary / important they are)

We should export the frequency data for the stoichiometry data and load that into yEd and see what see.

We should write some type of data loader that lets load in "hint-actions" so we can see where students request hints.

I should develop a Data-Properties class to more appropriately manage the different functions we have available depending on what type of data is read in. Basically the object just stores a dozen or so flags that are set based on what columns are read in, in the data-import stage of the program.

August 20th - 26th

Week's Plan:
Do a literature review on modelling problem solving.
Do a literature review on sequence mining.
Do some writing for graph-layouts, proposing the idea, edit the chapter on sequence mining, based on my findings for this week.

Week's Accomplishments:
I read a number of works on sequence mining and became more familiar with their goals, focus, and nomenclature. I also read up on modelling problem solving, which went much further away from logging tutors and states of problems and much more towards the state of the mind and ACT-R. It seems there is a significant gap between the states we can generate from tutors and data and the states modeled in psychology research. 

In the case of sequence mining that challenge is the very act of identifying the sequences. The problems have long target strings and really long data-strings, in the thousands of characters. In our case the challenge isn't discovering the sequences, though the algorithms developed from that field would be useful in our tool. Our challenge is to identify which sequences are most meaningful. For this, we need to look at the community structure and discoveries we can make from the networking field. By investigating the graph-invariants of the problem-solution space we can gain insight to which states are important.

I also wrote up my arguments and proposed ideas for investigating these aspects mentioned above.

Problems:
No problems these week.

Next Week's Plan:
Meet with Tiffany on Monday and discuss my proposed arguments. Continue writing up my proposal in a more concrete manner. I need to organize the text I have into a single consistent document. Where all the related works from each of the papers appear in the same related work section, combine the introductions, etc. and remove redundancy in the text. Specifically I will focus on the proposed research section and provide the necessary related works for those sections. I will have some text for proof-reading at the end of the week.

Other Pieces of Work: (I just don't want to forget these ideas, not sure how necessary / important they are)
We should export the frequency data for the stoichiometry data and load that into yEd and see what see.

We should write some type of data loader that lets load in "hint-actions" so we can see where students request hints.

I should develop a Data-Properties class to more appropriately manage the different functions we have available depending on what type of data is read in. Basically the object just stores a dozen or so flags that are set based on what columns are read in, in the data-import stage of the program.

August 12th - 19th

Week's Plan:
Finish step-based sequence work for presenting to Dr. Barnes. Discuss what to do next.

Week's Accomplishments:
Finished the step-based sequence work.

Problems:
The trivial case of calculating all sequences did not provide useful feedback. Naturally we need a more intelligent way of identifying important sequences, much like identifying important states.

Next Week's Plan:

Do a literature review on modelling problem solving.
Do a literature review on sequence mining.
Write a chapter for for graph-layouts, proposing the idea.
Edit the chapter on sequence mining, based on my findings for this week.


Other Pieces of Work: (I just don't want to forget these ideas, not sure how necessary / important they are)
We should export the frequency data for the stoichiometry data and load that into yEd and see what see.

We should write some type of data loader that lets load in "hint-actions" so we can see where students request hints.

I should develop a Data-Properties class to more appropriately manage the different functions we have available depending on what type of data is read in. Basically the object just stores a dozen or so flags that are set based on what columns are read in, in the data-import stage of the program.

Hours Worked:
Unrecorded, a bunch.