http://openresearch.baidu.com/u/cms/www/201210/30144944cqmu.pdf


On Mon, Jul 29, 2013 at 7:27 PM, Peter Muchemi <pmuchemi@gmail.com> wrote:
Apache guys seem to have something also

http://mahout.apache.org/




On Sun, Jul 28, 2013 at 9:13 PM, Adam Nelson <adam@varud.com> wrote:
Python ftw.

R is great but doesn't really have the toolkit integration for actual applications as such - it's more for statistical exploration.

---
Kili.io - OpenStack for Africa: https://angel.co/kili-io


On Sun, Jul 28, 2013 at 9:07 PM, Kennedy Kairu Kariuki <kkairu@gmail.com> wrote:
@Karunyu,

You'd need to build logic around the data collected as indicated by Martin. Human intervention comes into place also to filter out false positives. Its the same method used by research companies to monitor users collecting data in the field to ensure they collect data within their assigned location.



Regards,

Kennedy KK

Mobile: +254751110168
Skype: k.kairu
Gtalk: kkairu


On Sun, Jul 28, 2013 at 5:38 PM, Martin Akolo Chiteri <martin.chiteri@gmail.com> wrote:
Hello @Karunyu


On Sun, Jul 28, 2013 at 4:41 PM, Peter Karunyu <pkarunyu@gmail.com> wrote:
@Martin, that is an interesting approach, never heard of it before. How then would the software remember that certain drivers are more prone to going out of shape, or even co-relate data from several drivers so that if all of them go out of shape within a given window of time, that would not be flagged as mischief.

You simply store the data on the drivers' trajectories to particular destinations in a persistent form or storage. Relational databases are the most commonly used means for their simplicity and power. 

The frequency with which this collection of vehicle co-ordinates varies, say every five seconds, maybe more / less. This has interesting effects on what you can see on display or the kind of analytics that are made possible. At the end of the day, you need an individual to do the monitoring of situations surrounding the drivers. Probably you can find a suitable alternative to that. 

Martin.
 
@Peter, thanks for that document, it makes VERY exciting reading.


On Sun, Jul 28, 2013 at 8:51 AM, Martin Akolo Chiteri <martin.chiteri@gmail.com> wrote:
@Peter, 

Another thing that I have seen done in practice is that you can define the shapes of areas that the cab drivers should drive within as polygons on a map drawn to a certain scale in a system. After that just collect the co-ordinates of the taxis as they move in towns / cities /counties in real-time.

If they go out of the boundaries agreed upon, have the system raise an exception. If you have the driver's details, call them up and ask them what is going on. At times, their ``mischief'' is normally out of good reasons such as the presence of jams in road traffic or the occurrence of an unusual situation, such as an ambulance or a fleet of fire-fighters moving around.

Martin.


On Sat, Jul 27, 2013 at 8:43 PM, Kennedy Kairu Kariuki <kkairu@gmail.com> wrote:
@muchemi, how about using R? Though am talking from a theoretical point...



Regards,

Kennedy KK

Skype: k.kairu
Gtalk: kkairu


On Sat, Jul 27, 2013 at 3:16 PM, Peter Muchemi <pmuchemi@gmail.com> wrote:


On Sat, Jul 27, 2013 at 12:28 PM, Peter Karunyu <pkarunyu@gmail.com> wrote:
@Peter, now you have my attention :-)

Say, from a logical perspective, what constitutes mischief?


On Sat, Jul 27, 2013 at 9:02 AM, Martin Akolo Chiteri <martin.chiteri@gmail.com> wrote:
Hello Peter, 

Have you considered using Python and its libraries for performing Numerical [1] and Scientific [2] analysis? 

It also has tools for doing machine learning [3] with. 

[1] Numpy < http://www.numpy.org/ > 
[3] SciKit-Learn < http://scikit-learn.org/stable/ >  

They are all pretty simple to learn and use from my limited experience with them.

Martin.



On Fri, Jul 26, 2013 at 6:28 PM, Peter Muchemi <pmuchemi@gmail.com> wrote:
I consult for a taxi company that uses the stoic system to track movements of cars on a 24/7 basis.

They have a department just to pore through the stoic data to detect drivers mischief.

I have been considering deploying an expert system maybe done with prolog(using my 13 years old cs theory).

Anyone who has tackled such a problem before.

--
Peter Muchemi
Software Developer
Electra Software Ltd
http://www.software.co.ke
+254 722 320986



_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke


_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke



--
Regards,
Peter Karunyu
-------------------

_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke



--
Peter Muchemi
Software Developer
Electra Software Ltd
http://www.software.co.ke
+254 722 320986



_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke


_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke


_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke



--
Regards,
Peter Karunyu
-------------------

_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke


_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke


_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke


_______________________________________________
skunkworks mailing list
skunkworks@lists.my.co.ke
------------
List info, subscribe/unsubscribe
http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks
------------

Skunkworks Rules
http://my.co.ke/phpbb/viewtopic.php?f=24&t=94
------------
Other services @ http://my.co.ke



--
Peter Muchemi
Software Developer
Electra Software Ltd
http://www.software.co.ke
+254 722 320986





--
Peter Muchemi
Software Developer
Electra Software Ltd
http://www.software.co.ke
+254 722 320986