I managed to study quite a bit on the weekend.

I managed to read sections 4.1 and 4.2 in the book Knowledge Representation and Reasoning. Here I learned resolution of conjunctive normal form (CNF) expressions in symbolic logic. This covered universals, existentials, equality and variables. So I learned Answer Extraction and Skolemization. I can understand this but not sure I can do this myself so will need to practice the exercises in this chapter.

I will need in the next week and a half to review chapter 3 so I can begin to create a Knowledge Base (KB) for a small department inside a firm. This is one assignment I have due besides the lecture.

I also read more about basic knowledge management. I have started to find the book The Computerization of Work useful and interesting. It is not the sociology that will be useful but rather the case study descriptions of the context of working that will help me write my KB for the small department. This book is also critical of systems designers which helps me foresee some problems in my work.

I have to write a lecture due May 29th.

I am writing a lecture on knowledge loss, gain and retention for my professor and it is due May 29th. I have it started in $Latex$.

Finally success with creating a clean population data file

I was finally able to get a good clean data set for the world's population data using a different approach. In the original data set there was always a blank row at the end of each yearly table and at the beginning. Between these blank rows was a row with only the table title in column one. This title was a combination of the country name then a slash then the year covered by the table. My easy process was to label each row with the year as another variable. I did this by physically typing the year in the new column beside the title row then filling down in Excel to the blank row at the end of the table. Once I had the first country in this case actually the world tables with the new variable labeling each row in the cooresponding table with the tables year 1996 to 2050, I then copied the year variable to the next country Afghanistan. Then I highlighted both of the year variables for the World and Afghanistan and copied the years to the third and fourth country. Then repeated for the fitth, sixth, seventh and eighth and so on always doubling. This did not take long and it worked because the tables were all repeated with exactly the same number of rows per year table and per country.

The next step was done with brute force copy and paste so took all night but actually I did others tasks on the computer. It really took about as many copy and drag fills as there are countries. I highlighted the table title cell, the column one cell in the title row and then copied out only the country name and pasted this in a second new variable column and then I drag filled this down from 1996 to 2050 so each row concerning one country now had a new variable named Country that contained the name of the country. I then added column headers for variables names, Age_band, Both_genders, Male, Female, Year, Country. I then saved tihs excel file as a CSV file. Then I imported this to SAS. The using a data step where statement I deleted all the blank rows and title rows. So now I have this basic data file.

Knowledge representation and reasoning

The subject I am leanring from my supervisor is knowledge representation and reasoning. This is part of the artifical intelligence field. I have started to read in this field and have three books borrowed now. One of the books is written by my supervisor's thesis supervisor from the University of Toronto. This field is about logic in a sense that some forms of logic are knowledge representation languages. I am told I must choose a language to represent knowledge. I am reminded of a popular book that covered this idea of a knowledge representation and my whole topic of knowledge simulated in a computer. The book is:

Davis, Martin. The Universal Computer: The Road from Leibniz to Turing (New York: W. W. Norton and Company, 2000).

I am starting to write literature review sections.

I have the most reading done on knowledge loss in firms. So from this in brain reading I started the section of the literature review covering knowledge loss. I frame it as three issues, safety, cost of production, and intellectual capital. I am trying to construct ordinary academic writing that is at heart simple clear and not too articulate. I do not want to use the nuances of the social sciences in this thesis. I want to in effect use a style that is almost a lifting of structure but in effect I am seeking common structures of academic writing and even just a standard English prose. I will have plenty of fun and more original and arcane structure, when I start to document the computer systems and systems modeling. This, the results and methods sections, will also likely be that more difficult and most likely a very unclear part of the thesis.

I need to pick up the pace on my thesis.

I need to pick up the pace on my reading. May be I can finish reading a couple of chapters today as it is Saturday.

Narrative in thesis, music and aborginal studies.

In my early research into knowledge transfer the idea of story telling and narrative came up. But how do I add narrative to computer code? Narrative is a way older workers can relate their knowledge to younger workers. I was just reading the introduction to a book about an epic story. The book is

Van Deusen, Kira. Kiviuq: An Inuit Hero and His Siberian Cousins (Montreal Que.: McGill-Queens University, 2009).

The book contains an Inuit legend and compares this legend with Siberian aboriginal legends. This is how stories and lessons of life would be passed down in Inuit culture.

I can not spend time on cross culture too much in my thesis. This book is hundreds of pages long. My thesis will not be of use outside of capitalist firms. I thus can not read this book. As well the legend itself will be of no help rather the context of story telling might be of help.

I had also thought to read this book to help my music making but the book and story are not for telling thus to avoid plagiarism I will not read this book. I will simply be inspired by Kira's introduction where she explains briefly that she uses performance art, and music with story telling. She also explains that the elders who told her Kiviuq had given her alone permission to tell this epic. I like these ethics and will respect them.

I also thought, as I read the book, of aboriginal studies but have only read a few aboriginal studies books and will not read this one.

I started to read another book for my thesis today.

I read section 1.1 in

Levesque, Hector J. and Lakemeyer, Gerhard. The Logic of Knowledge Bases (Cambridge, Mass.: MIT, 2000).

This covers an introduction to knowledge bases but mostly an introduction to logic

I am reading the first book in my literature review and it is a good choice for my topic.

I read chapter one in

Linda Argote. Organizational Learning: Creating, Retaining and Transferring Knowledge (New York: Kluwer Academic, 1999).

Argote's research has been on learning curves. These are plots with hours of labour as a dependent variable v cumulative production as the independent x variable. The basic formula for a learning curve is as so: $y = ax^{-b}$. Cumulative production serves as a proxy for knowledge gained by experience in this model.

My paper on global demographics got kicked out of the social economy conference.

The organizer was not happy that my paper did not connect with the social economy so I withdrew my paper at his request.

I saw my thesis Supervisor yesterday and have not slept since.

I saw my thesis supervisor and he suggested rewriting the section titles of the first three chapter of my thesis. I have not slept because I have been watching hockey, playing computer games and doing work on my home computer. Well enough of all that I will now write my thesis. I have not been reading much these days as I am engulfed in work. I will now take a break from work at home and write my thesis. I must move to my eMac in my office now.

I joined the International Association for Statistical Computing.

Here is the mission statement of the International Association for Statistical Computing.


It is the mission of the IASC to link traditional statistical methodology, modern computer technology, and the knowledge of domain experts in order to convert data into information and knowledge.

Statistical Computing is meant here in its broad sense as Statistics in the Communication and Computer Age and it contains a rich variety of research topics in almost every branch of statistical inquiry.

The topics focused on by the IASC include computational statistics, statistical software, exploratory data analysis, data mining, pattern recognition, statistical graphics and data visualisation, statistical data bases, and related fields.

The IASC works to share and advance methods and techniques related to these topics by facilitating the exchange of information and by stimulating evaluations, research, and development in all areas of statistical computing. It is the aim of the IASC to foster world-wide interest in effective statistical computing and to exchange technical knowledge between researchers world-wide.

IASC promotes collaborative efforts within international, national, regional and other organizations and institutions having similar aims.

I am going to ProQuest, IEEE and ACM databases again for ten more items for background literature..

I feel I am done now searching out the knowledge management items for my thesis. I hope the next ten items are easier as I plan to search out simulations, programming and other methods literature now.

I can read in the country data sets in two lines of code.

I figured out how to read in the country data sets in two lines of code. The problem is I need 228 of these pairs of lines. To do this I turn to the automated features of "fill down" in Excel. I am learning that in SAS one can do this with macros. In Excel I have created the first line of code where I set a specific counter based on the position of the countries data in the large file. Once this so called counter or index of numbers is set the second line is a as.data.frame line that reads the large file but only reads the rows( as indicated by the counter/index set of numbers) needed for that country. So today I am going to use Excel to write the second line of code 228 times.

I am still using Inference for R and was mistaken in a sexist way about the representative changing her last name. Of course, many women have two last names and that was all it was. I have only completed the first tutorial for Inference for R and have not yet tried the studio application. This promises to allow breaks while running code.

At this point in testing this Inference for R software, after I have the second lines of code completed I will run this and try printing the Inference for R word document and adding a write up. I may use some of this blog for the write up. I will do this in word. I will after that try to make a Power Point for the conference and start to write up this part of the research. Ideally I will have only a few slides devoted to this part of the project but that show some of the code.

Then it is off to Google or back to the US census site to find data about major industries by country.

Beautiful quote from the Rae years by Lilith Finkler from Candian Woman Studies

Can you guess who her populations is, and why she was not given full academic rights at this conference even thought she was qualified academically but because she belonged to the subject population she was treated differently?

The politics of appropriation are such that a privileged few manage to determine subjects of study and gain recognition as "experts" on a community outside their own. This process becomes particularly insidious when so - called "progressive" individuals claim to be allies of an oppressed group and in so doing, constantly speak on their behalf. This disempowers the people directly affected by a particular experience, since they have less "credibility" and less access to privilege.

Finkler, Lilith. Canadian Woman Studies. Downsview: Summer 1993. Vol. 13, Iss. 4; pg. 72.

Inference for R software. Quality issues explored

I won't post something technical about inference for R here. The representative who contacted me from that company has now changed her last name in emails. And one search at Google found this debate from some noted statisticians about Inference for R as a spam marketed software. I was worried because I was approached outside of work. I am now finding that persons who contact me here about math have questionable ethics in terms of robbing Open Source projects. But this could just be theft. I may not be able to as careful in the future. May be Inference for R is a more nefarious product. I will need to search further and perhaps contact a computer security service now and inquire about this company and software. Perhaps it will all turn out to be an aggressive slightly unethical marketing side to a good software.

Practically twenty books or hundreds of journals articles will not work.

The two master's thesis advisers I saw who did not take me on as a student both offered the advice of not doing too much work on the thesis. I am realising now that twenty books would be too much. So just like a short list of twenty references is better than hundreds of references, a twenty item list with journal articles is also much more practical in terms of getting it done. Also books are not as easy to vet I think at present. Perhaps there are simple vetting processes for books but I am not able to know or express these simple processes right now. So I am cutting books down on the list and adding more peer reviewed papers. I am also focusing less on the seven components of operational knowledge and instead on knowledge loss and knowledge retention. Much better and seems like it will be much easier.

My searching of ProQuest that started a year ago is now filtered down to 79 references.

I am attempting to find the 20 classic references to the problem of Knowledge Loss. I may have nine or ten references now and these are books. I should be adding peer reviewed journals. I should cut some of these books and add more journal articles. I have only one or two classic knowledge management books and these two I will keep. And I have two methods books. And I have two knowledge loss books. I am looking right now at the possibility of narrowing down the 79 present references. I also now have the idea that I should perform a search on ProQuest and Google scholar for the term Knowledge Loss and find about 15 papers that way. After all, the previous searching that is now narrowed to 79 papers, was about searching for ways to measure knowledge. What I really need is papers on the main topic: Knowledge loss. Also I think ten books is too many books. So I will cut the book list to the six books mentioned above and cut the previous search to five or so papers then find ten or so peer reviewed papers on Knowledge Loss.

I am co-presenting a paper at a work conference.

The paper is about online seniors. Here are the conference details
Socio-economic Conference 2009

Baby-boom research

The conference I am presenting at is sponsored by this network The Social Economy Student Network (SESN) http://socialeconomy.info/en/english