Différences

Ci-dessous, les différences entre deux révisions de la page.

--- projets:plim:20142015:gr2 [2014/11/23 22:38] – bouillet
+++ projets:plim:20142015:gr2 [2014/11/23 22:53] (Version actuelle) – bouillet
@@ Ligne 42: / Ligne 42: @@
 The calculation of the speed of the user is made by calculating the distance between the last two points, and dividing
-the distance by the time elapsed between each data reception.
+the distance by the time elapsed between each data reception.\\
 Determining the most visited places by the user is more difficult. To enable this feature on our application,
-we used
+we used the k-means algorithm included in the library alglib.\\
+The major problem we've encountered is that it is impossible to calculate directly the k-means because the data we have
+is in specific GPS coordinate, named WGS-84. The k-means algorithm implemented required the data to be compatible
+with euclidean distance.\\
+To solve this issue the application converts on the fly the WGS-84 coordinates into euclidean coordinates, then
+the k-means algorithm computes clusters based on the user configuration. As soon as the clusters are computed,
+the application converts the data back into WGS-84 coordinates, which allows us to draw the clusters as green circles
+onto the map.
 === SOFTWARE PACKAGES of the Project ===
@@ Ligne 78: / Ligne 87: @@
 === RESULTS  ===
-The application works pretty well and give data sufficiently precise to get good information about the transportation method.
+The application works pretty well and give data precise enough to get good information about the transportation methods.\\
+Of course because of the GPS potential inaccuracies, sometimes the calculated speed doesn't match the reality.\\
 The clusters works pretty well too and groups position with accuracy to give a good preview of the visited location.