Différences

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

--- projets:plim:20152016:gr3 [2015/11/12 14:05] – benni
+++ projets:plim:20152016:gr3 [2015/11/17 09:23] (Version actuelle) – [SOFTWARE PACKAGES of the Project] simond
@@ Ligne 45: / Ligne 45: @@
 ==== Screenshots and GUI description ====
+<TODO>
 ==== SOFTWARE PACKAGES of the Project ====
+<TODO>
 ==== Required SOFTWARE Environment ====
+You need to have previously installed Visual Studio 2013 (Community edition will do the trick). The zipped solution will be directly importable from VS. You will also need Windows Phone 8.1 emulator to test our app.
+We tested it on Windows 10 but it should work on Windows 8.1 and 8.0 platforms.
+To build it we used Windows Phone 8.1 Silverlight project so you need to have WP8.1 SDK installed if you want to change anything in the source code.
 ==== How To build solution from project sources ====
+Import the solution sources into Visual Studio (File -> Open -> Open File -> open the solution file).
+From now, you just have to click on "build solution" and launch it. If you selected to deploy the app on an emulator, it will open the emulator for the specified device.
 ===== HOW TO USE IT and RESULTS =====
+==== How to use it ====
+The first time you start the app, it will ask you if you allow tracking your GPS position.
+Please say yes. If you answer "no", the app is just USELESS.
+Next screen displays the main screen of the app :
+on the top right corner, you have a map showing your current location
+on the top left corner it shows you your current latitude and longitude and cluster related
+on the bottom on the screen, it displays the different locations (Home / Work / Sport) and time spent on each location (i.e. on each cluster).
+==== Results =====
+  * Clustering works. We define 3 coherent clusters.
+  * The time spent on each cluster works.
+  * **But label put on each cluster is completly wrong.**
+Since the algorithm used to build clusters can change indexes for clusters built, cluster 0 will be "HOME" and next time cluster 0 will be "WORK".
+So it has a random behavior. To fix it, we should do the following :
+build 3 clusters from collected GPS locations (done already), label each location on each cluster with "HOME" or "WORK" or "SPORT".
+Next, we should count how many "HOME" location are tagged in cluster 0, cluster 1 and cluster 2. The cluster that has the most labels "HOME" should be declared as "Cluster HOME". The same logic should be applied to "WORK" and "SPORT". We try to do it but not achieved it (17/11/15).
 ==== Video ====
 ==== Tutorial ====