##INTRODUCTION## This is a solution for wlan positioning. It is proposed on the basis of machine learning algorithm and therefore composed of two seperate phases: offline and online. This solution is based on client-server mode.
Offline phase would process raw data, split the whole dataset into training set(70%) and testing set(30%), and eventually yeild a training model according to records from training set, using certain algorithm.
Online phase would launch a positioning server after some initialization, the server would wait for request and call predict algorithm to return a result.
Apart from offline and online phase, we supply an extra testing phase for users to examine both efficiency and accuracy of certain algorithm.
##HOW TO USE##
-
offline phase:
./offline.sh <device_dataset> <algorithm>.example: ./offline.sh htc NN
-
online phase:
./online -
testing phase:
./test.py <device_dataset>example: ./test.py htc
./gen_test.py <device_dataset>will randomly generate testing dataset from specified device_dataset
##DEPENDENCY##
scikit-learn is required
currently only available on linux(will migrate to windows soon)
##MANIFEST##
- online.sh
- offline.sh
- raw_data: measured data categried by device are available here
- alg: all algorithms available for training and predicting
- scripts: some necessary scripts, including the positioning server
- test: scripts for test are available here
##BUG REPORT LIST##
##WANTED##
Call for contributors.
This is a newly founded project. It's not 'what happened', but 'what's happening'.
Any advice or contribution are urgently wanted.