Comments on: Wolfram Mathematica Home Edition

By: un1crom

un1crom — Wed, 16 Dec 2009 19:34:58 +0000

so why don’t you and I work on that together?

By: SG

SG — Wed, 16 Dec 2009 04:02:35 +0000

There’s no question that you can do a broader range of things wtih Mathematica than any other software package on the planet. One of my favorite packages I wrote takes a set of urls (either files on your computer or web addresses), parses the text, and gives you all the standard measures of writing sophistication. Here’s the grades report for wiki.ergodics.org: {{“FleschReadingEase”, 61.9832}, {“FleschKincaidGradeLevel”,
7.62292}, {“GunningFogIndex”, 4.91041}}

It would be trivial if I were so inclined to have this thing crawl the net and rank the sophistication of say, the top 100 websites or perhaps find the most sophisticated authors on wikipedia.

To be able to do anything with data for $300 is a serious democratizer, because prior to Home Edition data analysis was only for professionals.

By: Mike Blaszczak

Mike Blaszczak — Thu, 12 Mar 2009 20:45:24 +0000

I think a rapidly growing number of users at home run a 64-bit OS. After all, four gigs of memory is less than $50.

According to Wolfram support, Mathematica Home runs fine as a 32-bit process on a 64-bit OS. That means it works, but doesn’t take advantage of the large address space.

Hoping that they correctly answered me, I’ve ordered a copy. If it works, I’m going to be tickled pink. As the Wolfram website says, it’s awesome that this version is available for us to work at home and just “play”; armchair mathematicians and casual non-student users alike are going to be thrilled with this.

By: un1crom

un1crom — Sun, 15 Feb 2009 21:30:40 +0000

Yes, though I don’t think that’s a “cripple” as most home users (a) aren’t on a 64 bit os/machine (b) aren’t going to benefit much from 64 bit unless they are doing massive computations, which would indicate they might want more than a Home Edition.

but, yes, your are technically correct.

By: Kursty

Kursty — Sun, 15 Feb 2009 13:25:59 +0000

This is NOT a fully functional version, as it is crippled to be only 32-bit on Windows, not 64-bit as the professional version.

Kursty

By: un1crom

un1crom — Tue, 10 Feb 2009 13:53:32 +0000

First thing is to get a set of images you want to search over

SetDirectory[
"/Users/russellas/Documents/NKS/imagerecognition/shoes"];(*set \
directory of images*)

imageList = FileNames["2_*"];(*get files into array*)

images = Map[Import, imageList];(*import the data into the notebook*)

colorMeans = Map[Mean[Mean[ImageData[#]]] &, images];
nf = Nearest[colorMeans -> images];

similarityGraph[n_] :=

GraphPlot[
Flatten[Table[
Thread[images[[i]] -> nf[colorMeans[[i]], n]], {i,
Length[images]}]],
VertexRenderingFunction -> (Inset[#2, #, Center, .5] &),
SelfLoopStyle -> None, ImageSize -> 3000];

similarityGraph[5](*Spit out a graph by color clustering*)

(*try it with outlines*)
imagesOutlines =
Map[Dilation[ImageAdjust[LaplacianFilter[#, 1], {.7, .6}], 0] &,
images];

outlineMeans = Map[Mean[Mean[ImageData[#]]] &, imagesOutlines];
nff = Nearest[outlineMeans -> imagesOutlines];

similarityGraphOutline[n_] :=

GraphPlot[
Flatten[Table[
Thread[imagesOutlines[[i]] -> nff[outlineMeans[[i]], n]], {i,
Length[imagesOutlines]}]],
VertexRenderingFunction -> (Inset[#2, #, Center, .5] &),
SelfLoopStyle -> None, ImageSize -> 3000];

(*show the graph*)
similarityGraphOutline[3]

now, to finish the search just put an image up against it and cluster it by color and outline.

DONE.

This was based on wolfram research image recognition samples in the blog.wolfram.com

By: chris willis

chris willis — Mon, 09 Feb 2009 18:42:26 +0000

You make some excellent points. Mathematica is viewed by some as something to be used purely for academic or aesthetic pursuits but it’s more than that – it a powerful sense-making utility.

And who doesn’t need more of that.

Count me in as someone interested in seeing your search code.