Différences

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

--- projets:plim:20142015:gr17 [2014/11/23 01:05] – ancienne révision (2014/11/23 02:00) restaurée ouhichi
+++ projets:plim:20142015:gr17 [2014/11/23 22:55] (Version actuelle) – [MAIN INFORMATIONS about the Project] ouhichi
@@ Ligne 46: / Ligne 46: @@
 ==  I- Project Description : ==
-Born from the idea of creating an application which makes a classification of persons per profile whether on foot, by bicycle or by car.
+Born from the idea of creating an application which makes a classification of persons per profile whether on foot, by bike or by car.
-Speed application provides users with their current location,speed and other informations and allows them a grouping oh their visited places.
+Speed application provides users with their current location,speed and other informations and allows them a grouping of their visited places.
 == II- Technologies and Tools : ==
+In this part we will define  some technologies and tools used in our project such us GPS sensor,K-means algorithm and local data base.
 **1- GPS (GLOBAL POSITIONING SYSTEM) :**
@@ Ligne 55: / Ligne 56: @@
   * **What is a GPS Sensor ?**
-In our application and in order to response to all our needs we used the GPS Sensor
+In our application and in order to response to all our needs we used the GPS Sensor.
-The GPS is a system of radio navigation by satellite developed and exploited  by the Ministry of Defence in the United States. it allows the users to determine their position, their speed and the hour of day in earth, atsea and in the air 24 hour / 24, in all weathers and to any place in the world.
+The GPS is a system of radio navigation by satellite developed and exploited  by the Ministry of Defence in the United States.
-The GPS signals are accessible by an unlimited number of users simultaneously and it’ s  given free of charge to all the users.
+The GPS signals are accessible by an unlimited number of users simultaneously and it's  given free of charge to all the users.
 Due to GPS sensor, you can get current information about users such as :
@@ Ligne 64: / Ligne 65: @@
 •	Longitude
 •	Altitude
-•	Time of fix (the time the reading is taken)
+•	Status
-•	Status (when currently receiving data)
+•	Speed
-•	NumSats (the number of satellite links)
+•	Bearing
-•	Bearing (the current direction or heading)
-•	Speed
-We can say also that GPS Sensor have some options like displaying it automatciclly when it s closed  by selecting "Show GPS Sensor" when GPS pane isclosed check box or to close it and showing the GPS pane.
+We can say also that GPS Sensor have some options like displaying it automatciclly when it's closed  by selecting "Show GPS Sensor" when GPS pane is closed check box or to close it and showing the GPS pane.
@@ Ligne 86: / Ligne 84: @@
 **2- K-means clustering :**
   * **The objective of the method :**
-Clustering is the process of partitioning a group of data points into a small number of clusters and it is one of the simplest unsupervised learning algorithms that solve the well known clustering problem
-The algorithm of the k-averages (or K-means ) is an algorithm of partitionnement of data raising statistics and machine learning. It is a method which divide observations K partitions(clusters ) in which every observation belongs to the partition with the closest average. The dynamic thick clouds are a generalization of this base, for which every partition is represented by a core which can be more complex than an average. We add also that the classic algorithm of K-means is the same that the algorithm of quantification of Lloyd-Max.
+The algorithm of the k-averages (or K-means ) is an algorithm of partitionnement of data raising statistics and machine learning. this method divide observations K cluster in which every observation belongs to the partition with the closest average. The dynamic thick clouds are a generalization of this base, where every partition is represented by a core which can be more complex than an average. We add also that the classic algorithm of K-means have the same functionality as the algorithm of quantification of Lloyd-Max.
   * **Algorithm method :**
-In the method of "k-means", the choice of the initial centers is made on the basis of a random edition without put back of k individuals from the population to be classified. The partition of the classes is modified with every affectation of an individual i of I.
+In the method of "k-means", the choice of the initial centers is made on the basis of a random edition without put back of k individuals from the population to be classified.
+Entry: k the fixed number of groups
+START:
+Choose randomly the centers of the groups
+REPEAT
+i. Allocate every case to the closest group to its center
+ii. Recalculate the center of every group
+UNTIL (stabilization of the centers)
+OR (number of iterations =t)
+OR (stabilization of the total slowness of the population)
+END
   * **Stop condition of the algorithm : **
-• When two successive itérations lead to the same partition
+• When two successive itérations lead to the same partition.
-• when  we reach the max number of iterations
+• when  we reach the max number of iterations(already fixed).
+  * **The algorithm's limit : **
+The reading of this algorithm suggests the following remarks
+• The method of "k-means" is strongly bound among the k number of classes fixed at the begining. However the classification l classes with l> k can be widely different from the classification k classes.
+• The classification using this method depends on the choice of the initial centers.
 ** 3- Local data base :**
-To store and retrieve data in we made appeal to a local database.
+To store and retrieve data in our project we made appeal to a local database.
-This local DB uses uses LINQ to SQL which provides an object-oriented approach in order to work with data and comprises an object model and a runtime.
+This local DB uses LINQ to SQL to allow an object-oriented approach in order to work with data and comprises an object model and a runtime.
+== III- Project Step by Step : ==
+**1. Data Collection**
+The GPS Sensor used in our project allows the users to determine their position, their speed and the hour of day in earth, at sea and in the air 24/ 24 hour, in all weathers and to any place in the world.
+In the launch of the application, there will be a restoration and a recovery of all these data then posting them on the first interface «transportation mode”.
+**2. Data Storage**
+Once Data are collected it will be stored in a local data base.
+The local DB will be useful on one hand to apply various processing like classification and on the other hand to
+Keep a journalisation of the activities of the users and the possibility of displaying them at the needs.
+**3. Profiles classification**
+After collecting Data we apply a classification by speed on users in order to obtain finally 3 profiles whether on foot, by bike or by car.
+And this classification will be marked on our page with small images which signify each case.
+**4. Clustering**
+In our application once profiles are classified we ally K-means algorithm to have a clear distribution in form of group of the places the most visited and seen frequently by the users.
+These cluster will be shown on the maps and marked with circles.
-== III- Application functionnalities : ==
+== IV- Application functionalities : ==
@@ Ligne 147: / Ligne 202: @@
 The final result of our project is to have a clear classification of users according to their profiles as well as a classification by cluster of their busiest and most visited place.
+Our application emphasizes the importance of the non-supervised classification and presents a direct resolution of the recognition of the modality of the transport problem of our everyday life.
-=== Optionnal but considered in the evaluation : An 2 pages ACM paper on your results ====
+===  2 pages ACM paper on our results ====
-// Optionnal but considered in the evaluation //
-//if you finnd enough time, you can write a 2 pages paper about your project in a professional ACM format. This paper will be submitted to other researchers in the field and perhaps to present  a poster in a conference and/or Workshop
-//
 **A Paper ACM Format :**
-[[http://www.acm.org/sigs/publications/pubform.doc|Word ACM Format]]
+[[https://github.com/HAYKEL88/Speed-Clustering/blob/master/Vitess++/Assets/ACM_Word_Paper1.pdf?raw=true|Word ACM Format PDF Version ]]
+[[https://github.com/HAYKEL88/Speed-Clustering/blob/master/Vitess++/Assets/ACM_Word_Paper1.doc?raw=true|Word ACM Format DOC Version ]]