0
Posted on 9:43 PM by 4 8 15 16 23 42 and filed under


It is 22.56 pm and I've been trying to sleep for two hours. Why so early, you may ask. Because I've lied to myself and believed that I'd better wake up early and write my thesis.

This last past years, I've been alone with myself a quite whole lot more than before. I imagine one thought flowing to a stall sea - and slowly breaksdown the harmony. On one hand, I say to myself, you are the one who is different, you are the one who is promising. Do something, just not watch Youtube videos or stream previous shows you've already watched. Yet, there is this alter ego, maybe a secret identity, whose life goal is to constantly drown me down. It is maybe a mother figure, saying boy, don't worry - you'll work more tomorrow. Still it is hard to decide whether it is in fact an evil nemessis or a caring relative.

Can we say that, as in movies, do major periods in our lifes have themes? Primary school years: playing Power Rangers in recreation, naivite, simplicity. Middle school years: patience, waiting for it to finish. Highschool: peace, hope. College: Search of identity, trying to understand the dynamics. Master's: hope, patience, peace. PhD: Fall-down. Self-interrogation. Feeling of insufficiency. Desire for impact along with a feeling of incapability. What is next?


When are we honest with ourselves? When do we lie? If we put our self-protection mechanism on a side, would we still like what we saw? Why do we obsess on different things? And most importantly, why can't we control our thoughts?

Even if I know a particular way of thinking is flawed, I still can not stop it. Take this city obsession for example, I know that by fantasizing about different cities and not seeing any brightside on my current one, I am making a mistake of confusing fantasy with reality. When I dream about New York, I completely forget my earlier dreams about Austin. I know that this is hypocrite but why then I can't stop? Don't I know it well enough or does hope triumph experience?


0
Posted on 8:45 AM by 4 8 15 16 23 42 and filed under


Hello my fellow followers,

You probably think I am dead. No offense but I thought you disappeared as well. For a while, I thought although this blog was highly popular in many circles, it was a miracle that Google Analytics would skip any sign of traffic. This lousy blog I created, tried to imitate from others obviously did not work. Still, I ask myself are my thoughts really worth expressing? Yes, is my answer for now. No, you don't have to read it - purely because I have to get these out of my system, sure it could be nice if it would help some other in anyways but that is not my concern.


When I first got excited by the idea of writing a blog, I was under the influence of some bloggers that I followed. This reflected to my initial lousy posts which were fun for a while, then I got bored and stopped completely. It was useful for me to combine ideas from various books and motivated me to summarize them while organizing the arguments in a more clear setting. Summarizing books required some work, but I think its was worth it. The blog did not have any purpose, still does not. This post - and I guess this blog so - is meant to help me, here it is its purpose.


I moved from my beloved Istanbul to the heart of Texas (not quite, Austin is not that much Texas) - to pursue a PhD degree. Moving to States was something I wanted for years as I come from a generation who grew up to have "Back to the future" as their favorite movie, and U2 as their favorite band. I guess I would feel more of a connection to country music in Tennessee, then eastern Turkish folk music. Don't get me wrong, I love the land I grew up and I feel connected to Istanbul more than anywhere but still being imposed to that much American culture makes you feel like an insider even when you are outside.


Still I got surprised when I came here. Not because I realized it is obligatory to use the word "like" at least 5 times in a sentence, or you say thanks to the bus driver when you get off ( at least in Austin). I was more surprised by my own emotions and my understanding of the culture surrounds me. I thought I knew everything about the culture, and most of the time my experiences affirmed my initial impressions. However, I  started to question more the lifestyle  Maybe I had read Ayn Rand so much, but I still respect the idea of self-made man. Yet, the reflections of this idea to the culture are complicated. In this country, people are not disfortunate, they are seen as losers. If you can't achieve something it implies that you are a failure. Here I am exaggerating a bit but on a cultural level pure rationalism rejects all super-naturalism which in turn generates a culture that lacks a sense of contentedness. If you knew me you would laugh your ass off, as I act like logic is a new religion. Still, before observing at first hand, I never thought I'd miss little nuances that my own culture (or Europe) has, and America doesn't. Or I am just really homesick..


The fact that America lacks some qualities that Europe (+ Turkey) has, does not mean I regret my choice of coming here or that it is a superficial culture. I have always been interested in other cultures especially western ones, which motivated me to learn french and italian. I adore those cultures, mostly like those people and would like to spend years in Rome or Paris. I have good grasp of both languages but in any case I felt like I would feel like an outsider in those countries. I felt like no matter how many calzones along with vino rosso I would never feel like an italian. Not that it is important to feel italian, I think what is important is the feeling of an outsider. Wouldn't it be exhausting after a while? Yet, I thought I would never feel that way in States - after all anyone can be an american. I would express myself without any worries and I was already imposed with americana. One other feeling I got was - this may be totally wrong but note the word feeling - Europe was for fun, not work. I really liked having camambert, but I thought I could have that later on after having achievements. Where would give me more opportunities? My answer was America. This was a country where I could be a millonaire if I create an iphone app that makes fart noises. Or I could build a new type of McDonalds that serves to only lactose-intolerant people. Pardon me for the irony, but I have always felt that if you are good, this land would give you most opportunities - and I think the same way today.


My ramblings will continue shortly..




















0
Posted on 7:25 PM by 4 8 15 16 23 42 and filed under
Couple of notes from the debate:



THE TWO-PARTY SYSTEM IS MAKING AMERICA UNGOVERNABLE



I hope posting these would help me remember them:)


2-party system does not work
2 party system does not motivate young : 39% independent voters
Media does not help to propose any change
Parties impose certain ideas, do not allow flexibility to politicians
USA is built on the principle of competition: why not in politics?

2-party system works
 It works , brings stability
It still works, the problem is current environment not the current system
This is not a dictatorship, does not need social media to be changed there is always an alternative
Brought up the Obama movement



0
Posted on 7:32 PM by 4 8 15 16 23 42 and filed under ,



Since this Blog is about stuff as the title shows( btw I really should consider major changes in this blog. Come on what’s a blog about stuff, how cheesy is thatJ ) , it’s time to dedicate this post to “stuff” itself.  Or, maybe to its story.

“The Story about Stuff” is a documentary about the production-consumption and disposal process of every item that we use in our lives. The narrator Annie Leonard, a former Greenpeace employee with a envorinanmental studies degree, is the one who introduces us this story about what goes wrong in our production-consumption cycle and how we can improve it.


The 20-minute documentary is definetly worth of watching( Animations rule!). Video looks nice persuasive but I wonder if that story is actually the right one?

Actually, there are many critiques to the Story, a quick research in Google and I found interesting results

First, start with the good news. The new school of thinking that Leonard suggest, are mostly accepted without any hesitations. In other words, there’s nothing wrong with  “Green Chemistry, Zero Waste, Closed Loop Production, Renewable Energy, Local living Economies”.

Whereas the arguments of Leonard seem to attract more opposite views. For example:
“Not everything about manufacturing is “toxic.” Make no mistake, I think there are plenty of dangerous chemicals in the things we produce, but you’re playing fast-and-loose with reality if you’re going to say that manufacturing is the simple practice of putting toxic chemicals onto stuff to produce toxic products.”
“ Annie Leonard raised the freak-out level: We dip our pillows in BFR (brominated flame retardants), a horribly toxic man-made chemical, and we sleep on them!
Not knowing anything about BFRs, I did some research on this. I found two things:
One, BFRs are used primarily in electronics and electronics plastics. Things like computer circuit boards, the plastic casing around a TV set, around the rubber sheaths encasing wires in a computer, that kind of thing. And the BFRs are chemically bonded to those components. That means they’re not flying into the air, we’re not breathing them in.
Two, I was unable to find any evidence of companies dipping pillows or pillowcases in BFRs before selling them.”
There is even an Alternate Story of Stuff video series. It critiques Story of Stuff, on Youtube.

From my point of view, something bugs  me. This video cries out  “WE ARE IN CRISIS!”, it reminds me this quote from Michael Chricton:


“The media did  not make extensive usage of terms such as crisis, cataclysm,plague or disaster. For example during  1980’s, the word crisis appeared in news reports about as often as the word budget. In addition prior to 1989, adjectives such as dire, unprecedented, dreaded were not common in television reports or newspaper headlines. But then it  all changed.
These terms started to become more and more common. The word catastrophe was used five times more often in 1995 than 1985. Its use doubled again by the year 2000. And the stories changed too. There was a heightened emphasis on fear, worry, danger, uncertainty, panic. “

Long word short, media sells catastrophe news and people buy it. That’s the thing that annoys me, I really can’t decide now. Does Anne Leonard really tells us the truth or sells some part of it?

I think the main problem about the Story of Stuff video is that, it lacks detail and explanation. I know, it's a full 20 minutes, but she needed to explain her points more. I guess she tought so, too. As of March 2010 the book of Story about Stuff is available. I hope this time we may really understand the story about stuff. 





Arda

0
Posted on 4:36 PM by 4 8 15 16 23 42 and filed under

Again last semester, for my Statistical Inference Course I was supposed to perform Multiple Regression Analysis on a dataset. In this post, I will share my notes on finding the data and steps of the study and about software selection.

1.        Finding the Data:

At the beginning I thought  it’s a great oppurtunity to investigate Global Warming. In regression analysis, you try to explain one variable as the function of others and make inferences. So you have several variables namely predictor variables( the ones you use in prediction of the response variable) and one response variable( the one you are interested in).
Coming back to Global Warming example,  my initial idea was trying to predict temperature by using C02 level, deforrestation level etc. Why didn’t I do it? Maybe “searching for hours for these data and not finding as I wanted”  is the reason.
The challenging part in finding data was that for each different data the sources are different. You have to find different official data from different sources since they are not published together in most of the cases. I could find Temperature Data but I was close to finding CO2 but 2 more problems:

1.       There is need to organize the data / Time- Consuming + Time- Consuming+ Time- Consuming
2.       I needed more predictor variables (10 would be enough but I had only 2) since in the analysis you decide on whether to keep or omit some variables.

   So I gave up my idea on global warming which as a subject looked pretty interesting but finding data and organizing them seemed challenging thinking of the above problems. So what I did? Found a sample dataset.
Coming up with a data yourself, and organizing is not really the purpose of regression analysis. In most of the cases, you are supposed to apply regression techniques on a dataset, no matter what your dataset is.  Apparently there are statistics websites presenting datasets for statistical purposes. The great part about  that is they present you many sets and you can choose among them which one you are interested in.
http://lib.stat.cmu.edu/ is an example statistical website. (The one I found mydataset) The datasets presented here, are not made up to give nice results, instead they are taken from studies or research papers some time ago.
http://bus.utk.edu/Stat/DataMining/files.htm : This site also gives dataset which you may use for regression. I got the impression that Senic dataset is a very famous one and could be used without hesitation.
You may look at these websites also:

2.        Steps of Multiple Linear Regression Analysis


Test the Signifance of Regression : It’s a basic F test. You do test to check whether regression analysis is really helping you explain a significant amount of variance on the  response variable.   

Diagnostics and Remedial Measures  based on all predictor variables : Initially start with the full model where you put all your data. By looking at residual plots you decide whether assumptions below are violated or not, and if they are by transformations you try to correct them.
·         Multicolinearity Between Predictor Variables
·         Linearity of Error Terms
·         Constant Variance of Error Terms
·         Independence of Residuals
·         Normality of Residuals

Selection of Significant Predictor Variables: After you made sure that full model satisfies all the regression assumtions, in a whole set of available variables you take the ones which helps you analyze a  significant amount of variance on Y(response variable). You can either use Stepwise, Forward Selection or Backward Elimination in deciding which variables to include in your model.


 Diagnostics and Remedial Measures  based on Selected Predictor variables
·         Linearity of Error Terms              
·         Removing Outliers        
·         Constant Variance of Error Terms
·         Normality of Error Terms


Validation of the Model :  You initally save some of the observations and not include them in regression. Let’s say with 400 observations you calculated an apropriate regression line and  for  100 samples you look at how does your regression line fits.

3.       Which Software to Use:

Well, you can use many statistical softwares for regression but I guess  SPSS and Minitab are the most popular ones. I first started with SPSS thinking that it’s much more popular software and knowledge of Spss would be a plus because its widely used. At the beginning it looks simple, it is indeed but for some simple tasks Spss is just a waste of time.
For example you want to plot all the residual plots and copy them to Ms Word, Excel etc. , in Spss you have to copy manually each graph manually and it takes time. Editing Graphs,or plotting mutliple graphs in Spss is not easy either.
So I looked to Minitab and Result: Minitab Rules! You can plot variables against each other very easily and save them all in .jpg  files. Plus it seems to me that in Minitab you have more options in regression, so I would definitely suggest you to use Minitab. Also, trial version of Minitab  is available for 30 days( Way enough for regression projectJ ).  

I hope this blog post helps you in your analysis.

End of Course-Related Blog Post Series!!
{ at least for a while}

Arda
0
Posted on 5:04 AM by 4 8 15 16 23 42 and filed under

Last semester I took IE 501- Linear Programming and Extensions Course. At the beginning I thought the ‘Extensions’ would mean solving large-scale Lp problems in different areas, it turned out that the term ‘extensions’ is in the sense that you dig in the theory of linear programming. The course focused mostly on simplex algorithm in a much deeper way.  In undergrad, you learn how to implement this algorithm but in graduate study you learn to combine linear algebra and simplex, thus you study simplex in a much wider way. It seems like having studied simplex before would be a plus for this course, interestingly it does not a make a significant difference, because you look at it in a very different perspective.

1. Coding Simplex in Matlab

For a term project we were supposed to code simplex in Matlab and solve almost 100 BIG instances (maybe I should say giant).  In other words, create a Solver like linprog of Matlab. Our objective was solve all instances with our solver, correctly and fastly possible. The project group with lowest average Cpu time in solving instances would get the highest grade.
Matlab is easy to learn and there are many tutorials and websites available in internet in which you can possible find answers to almost all matlab related questions. In many cases, I think Matlab Help is the most clear and comprehensive source.

Coding simplex algorithm is not very very hard to do but as you cover all assumptions( degeneracy, full rank, stardization) it gets challenging.  It’s better to write at first the most basic  form and that make additions or changes to the code afterwards.


George B. Dantzig (1914-2005), inventor of the simplex method

One of my major loss of time was to use zillions of functions in the Code. I don’t mean the Matlab Built-in functions or anything, it’s just I separated my code in parts (function optimalitycheck, function calculate_new_solution). The simplex code is not very long, and although using functions for different operations seem like a neat way in debugging etc., it actually caused me to look at every function when there is a problem with the code. Waste of Time! So long word short, I’ve used 3 functions in the main code: (1) initial feasible solution, (2) basis update, (3) minimum/maximum ratio test .

Since we wanted our code to solve instances fast, we used Matlab built-in functions (like find, min etc.), which perform better/faster than the for loops that you could write. But for speeding up the code, the first thing you should check for are the unnecessary for loops. Other than that, there are some tricks like using logical indexing, but focus on ‘for loops’, they are the killer ones of your code.

2.Finding an Initial Dual Feasible Solution

The Simplex algorithm has 2 major methods of reaching to optimality. Primal simplex and Dual Simplex Method. In both cases, you have to find an initial solution and than execute second phases of the algorithms and find optimality/infeasibility/unboundedness. For primal simplex method, finding a primal feasible basis/solution is easy. The easiest way is to adding artificial constraints and removing them from the basis, iteration by iteration. Internet is full of sources on primal phase 1 algorithm. Whereas, finding a dual-feasible solution is different.

I spent DAYS to search on how to find a dual-feasible (a basis which is not primal feasible and assures primal optimality). There are websites which explaines how to find a suitable basis for bounded simplex( in which there are upper and lower bounds on variables) but I couldn’t find a single source showing how to do it for standart simplex.

****

A little break.

(I type all the possible keywords to attract a similar drowning person to the answer)
Dual Feasible Initial Basis/ Feasible Dual Initial Basis/ Initial Feasible Dual Basis/ Initial Dual Feasible Basis/ Dual Feasible Initial Solution/ Feasible Dual Initial Solution / Initial Feasible Dual Solution / Initial Dual Feasible Solution / Basis For Dual Simplex/ Dual Simplex Initial Basis

Sorry for the break. The show continues…

****

So we used artificial constraint technique described by Bazaraa. (Books still have stuff not found in Internet)







This method consists of basically adding an artificial constraint(2) to the original problem in which a primal feasible basis is available(can be calculated via phase1 primal simplex) and extending the basis. Then after one primal simplex iteration, the basis becomes primal infeasible and dual feasible. Aloha!

Here is the pseudo code of the algorithm:



If pseudo code above seem complex to you, you can always check out Bazaraa’s textbook where this technique is described with a numerical example.

This method is criticized in literature since the algorithm performance depends on a value of M and also it necessitates problem to be in the standard form. Converting a problem in a non standard form to the standard form results in increase in both the number of variables and the constraints (That’s a major drawback!). Yet, efficient methods exist which can accomplish this conversion with the least increase in the problem, the problem still enlarges. Therefore, many sources offer a different method for initial solution. They mostly focus on bounded simplex method but that was out of scope for our project and seemed more difficult.

That method was the most practical to implement for our case.

Finally, some notes on solving the BIG instances with our algorithm. Ok, our algorithm was working perfect with small instances (the ones we could revise by looking at iterations when we are not too lazy/ or shortly TOWCRBLAIWANTL) but when it comes to Giant instances sometimes our code couldn’t find a solution. Since they are not TOWCRBLAIWANTL, we couldn’t possible look at all iterations and find out where it screws up. I guess, the bigger the problem’s size is, or the more there are multiplications, there comes rounding errors, which result in absurd results. Also we thought it can do with the epsilon value we assigned, but blaming Matlab could be a suitable thought alsoJ.

Anyways, I hope these information would help someone and you don’t waste your time like I did.

I’d be glad if you leave a comment if you find this post useful.

For that someone should read this post except me.
                  
                   For that this blog should be discoverable

                                     fprintf(‘Hear this Google!’);

                   end
end


0
Posted on 4:50 AM by 4 8 15 16 23 42 and filed under
This January I was pretty busy with my course projects. In graduate level, the projects are generally more difficult than in undergrad and the projects are really detailed.

It’s not that I didn’t like the courses I took last semester but the time to focus on projects was insufficient, therefore I got really tired.  I spent so much time to reach some information and I remember typing similar keywords to google for hours and not getting useful information. So I thought I could summarize my projects and give some details which took me a lot of effort and time to reach.

Hopefully, someone who suffers the same problem as I did, may find these blogposts useful. (And this blog would really serve a purpose )

So I will write 2 blogposts and give details about my Linear Programming and Regression Projects. 



My knowledge is all yours.