I was rambling on the google discussion list when I found this example of how integration of online resources on a map visualization can help to solve lots of simple little problems in our daily life.

I just had a similar experience.  I wanted to search for the store  Meijer on Google Maps, but could not recall the correct spelling.  I  ended up doing a regular Google search on the wrong spelling, and wound  up with a lot of pages about computer game designer Sid Meier.  Then I remembered Google Sets, and entered “Wal-Mart, Target, Sam’s ?Club” as a set.  It immediately brought me back a list of those and ?other major stores, including “Meijer” (correctly spelled), and, ?interestingly, both “JC Penny” and “JC Penney”.  A quick trip back to Google Maps, and I found the nearest Meijer store, ?3.9 miles away from my office.  Thanks, Google!  — One Happy Googler

Tags: google, Location Based Services

F. Micheloud. Correspondence analysis. In Proceedings of the Advanced Econometrics, Lausanne, Switzerland, 1996. HEC. [url]

——————————–

This paper introduces to the idea of Correspondence Analysis, which is a factors analysis method that uses categorical variables (that is, non continuous or discretized ones).

The foundamental idea of this method is to eliminate the redundancy of the original data thanks to a reduced number of variables (the factors) which are a combination of the original variables.

Correspondence analysis is a method to describe synthetically a contingency table in which homogenuous individuals are classified on two criterias. This method has been developed by the french-lebanese Benzecri at the end of the 60’s for linguistic applications and is based on the notion of profile.

Tags: Exploratory Data Analysis, statistics, text data mining

Today I have been thinking a lot on the Algorithm that I have to implement for the core of my thesis. I reshaped the diagram that I used for my proposal to include the recent ideas that I developed while reading. This is a pretty raw draft. Several dimension are still missing:

1- a social rating mechanism;

2- a feedback system that should improve the tagging;

3- a clear distinction of what is a message and what is a landmark.

4- normalization and homogenization of the matrices.

5- maybe something more …

Tags: clustering, map algorithms, spatial clustering

The Porter stemming algorithm (or ‘Porter stemmer’) is a process for removing the commoner morphological and inflexional endings from words in English. Its main use is as part of a term normalisation process that is usually done when setting up Information Retrieval systems.

Tags: clustering, text data mining

(via)

A stemmer is a program or algorithm which determines the morphological root of a given inflected (or, sometimes, derived) word form — generally a written word form.

A stemmer for English, for example, should identify the string “cats” (and possibly “catlike”, “catty” etc.) as based on the root “cat”, and “stemmer”, “stemming”, “stemmed” as based on “stem”.

Tags: clustering, text data mining

A. Lagger. Distances sémantiques et extraction de thèmes. Semester project report, CRAFT – Ecole Polytechnique Fédérale de Lausanne, Ecoublens, Station 1, CH-1015 Lausanne, Switzerland, 2005.

———————–

This report present the semester work on the theme of semantic mining carried out at CRAFT. The goal was, given a certain corpus of texts, to define the semantic similarity of the text and extracting the relevant themes.

Two different algorithms have been implemented and tested. The first is the Condensation Clustering Value (CCV) described by Wise, 1999. The second was the Term Frequency / Inverse Document Frequency (TF/IDF). 

Tags: clustering, map algorithms, text data mining

The goal of this project is to decipher the world that news media reconfigures and to observe if media coverage, or lack thereof, is creating a new cartography.

Vanishing Point consists of a map of the world connected to a database fed by news coming from several international newspapers.  The visibility of each country on the map results from the quantity of media coverage the country receives, so those countries that do not make the news disappear progressively.

Tags: anamorphosis maps, information visualization, map algorithms, maps, politics, society

A. Kraskov, H. Stögbauer, R. G. Andrzejak, and P. Grassberger. Hierarchical clustering based on mutual information. Bioinformatics, pages 1–11, 2003. [url]

————————–

This paper presents a new method for hierarchical clustering of data called mutual information clustering (MIC). It uses the mutial information MI as a similarity measure.

The core idea is that the object to be clustered can be either single finite patterns or random variables. (i.e., probability distribuitions). In the latter case we can compute a similarity measure using the Pearson correlation coefficient.

According to the authors, this technique is not sensitive to nonlinear dependencies which do not manifest themselves in the covariance and therefore it can miss inportant features. On the contrary they propose the MI measure as a good indicator of similarities, saying that this parameter is zero only when the two random variables are strictly independent.

Tags: clustering, information metric, pattern recognition

This module is a wrapper for the Google Web APIs. It allows you to do Google searches, retrieve pages from the Google cache, and ask Google for spelling suggestions.

Tags: hack, google, Python

C. H. Bennett, P. Gács, M. Li, P. M. B. Vitányi, and W. H. Zurek. Information distance. IEEE Transactions on Information Theory, 44(4):1407–1423, July 1998. [url]

——————————-

This article contains the definitions of the Information Distance marker, as long as other related concept such as minimal/maximal correlation and Kolmogrov complexity. This ID is defined as the minimal quantity of information sufficient to translate between x and y, generating either string effectively from the other.

The algorithmic information distance is defined in the same way as the the  shortest binary program thatcommputes x from yas well as computing y from x. Being the shortest, such program should take advantage of any redoundancy between the information required to go from x to y and the information required to go from y to x.

Finally another intersting definition is that of the cognitive distance. The author refers to metrics as the Humming or the Euclidean distance. However, the author stress the fact that a small change of bits in the binary represenation of two pictures, for intance, reflects on a higher euclidean distance. On the contrary the manipulated pictures still will look the same. So they came our with another metrics that should take into account patter similarities. They assume that similarities between objects can be represented by effectively computable functions of binary strings. 

Tags: algorithmic information theory, pattern recognition, information metric, text data mining, universal cognitive distance