Register     Lost password?
Application details
Alfred, the emotional co-driver
Alfred the emotional co-driver reacts on your driving style. The driving style is recognized with machine learning technique (k-nearest neighbour classifier). For more information please see pdf file:


Please note this is just a "Alpha" version of this app. There is much which can be improved but today the contest ends :-(
App submitted: 30.06.2010
Social Networks:
Share this App on Facebook!
In order to evaluate this application, you will have to register and confirm your account.
@TeamCharlie ;-) Sorry I think you missunderstood me! I really like your apps... They have a very nice design and you have some good ideas! Nevertheless I do not find the similarity between DRIVEGUARD and ALFRED. User2858
Hello Lars, thanks for this comment. Yes - this is a presentation. Later, when the IDEA will (maybe) win, it will use all the CAN data you also mentioned. Best, Team Charlie PS.: Relax :D User3101
I don't think that there is much similarity to DRIVEGUARD. DRIVEGUARDis just a presentation with no functionality and Alfred detects driving style on CAN data with machine learning technique. User2858
Hi Lars, that's very similar to the cute drive guard...?! User3273
Hi Lars... Funny app :) User3040
A very good job with Alfred :)!! Sadly, as you said, today is the last day... I hope you finish it :) User2830
This is great! User3319
Thumb Up User2895
Member details
member since: 07.05.2010
Apps: 7
Ideas: 7
Messages: 47
Comments: 112
Evaluations: 93
Activity Counter: 953
For classifiers you have to have predefined classes, which I guess would be 'Ok', 'Angry' and 'Happy' Alfred. If you have SVM classifier, you would fit the current data points and see what the SVM classifies that point to. I guess that's the standard procedure we all follow. But I guess my concern lies with the classifier that we make, which I assume is using K mean approach. Using the dissimilarity matrix you choose the class to which its least dissimilar to. My previous message hints the error with deciding the classes, which by the way are formed from Training data points. But the inherent variability is too high, that you would end with so many wrong classifiers with the Linear classifiers that you plan to develop that it may give very absurd results. User3233
Hi VisteonIntern! Thanks for your critical comment and your interest on my app! I will try to explain you my thoughts :-) The first thing is i didn't use k-means which is a clustering technique which clusters a set of datapoints in a predefinded number of clusters and you are absolutely right that it is hard to determine the number of clusters before. Instead I used k-nearest neighbour classifier (I admit that the two names can esily be confound). Since I don't know if you are familiar with the concept of classifiers I will give you a short introduction: A classifier works as follows: 1. Step: trainig phase - in the training phase let me say you have to create the "world" with its elements the classifier should know. Let me give you an example: If you want to classify music acording to genre and you have four genres (classical, jazz, rock, rap), then you have to tell the classifier what are the characteristics of the different genres. So you will take many music clips (samples), calculate some features (beats, onsets, pitch, timbre ..) and train the classifier 2. Step: classification phase: now you will show the classifier a new music piece (sample) which he has never seen before. This peace will be compared with the peaces that were used in the training phase. According to the operation mode the classifier will then find the genre class to which the new peace belongs by mathematical calculations. In case of k-nearest neighbour the euclidean distance from the new sample to all samples from the training phase is calculated. According to the choosen number of "K" the k nearest samples from the training phase to the new sample are considered. If the new sample is mostly e.g. near class "classical" the classifier wil return class classical as result for the new sample. ---- now to your approach to use fuzzy logic. I think fuzzy logic is also a classification technique, but I must admit that I am not so familiar with it but I know that fuzzy logic is mostly used when there is no mathematical description of the "world" but a verbally like in speech recognition or automation of technical processes. -- I hope I could clarifiy your doubts a little bit ;-) if not feel free to ask me again or to tell me what you think is wrong with my approach! User2858
Hi , I don't mean to sound too pedantic, but I doubt if K mean Clustering will give you any reliable results! My personal choice would definitely be Fuzzy Logic for what you intend to do. K mean being sensitive to initial number of clusters you would choose, it will automatically try and reduce the in-cluster variability trying to reduce the inherent variability that the driver would have, hence leading too absurd results. I also cant logically convince myself , what distance matrix to use and why! If you don't mind please do share some technical details, your approach seems to be interesting, but flawed. User3233
Strange ... it is the 1. of Juli 7:37 and the contest is still running! I am confused! User2858
I've just seen. As I've written a very good job. A pity that the contest ends today. and I must say I really enjoyed both your applications and your comments and suggestions. I hope to see you in future contests:). Greetings! User2830
Hi my frnd lars, Today is the final day of this contest.i think we all miss each other from 2moro onwards.Last 2 months we stay here like a community.Anyway its time to leave.Thx for ur support & valuable comments. Wish u all the success & good luck.Hope we can keep in touch. Thx again my frnd User3148
Hi Lars! Thanks for your welcome :) I'm waiting your latest app that sevennine comments. good luck dude! ;D User2830
Hi Lars, good to hear you are working on another app :) the contest ends by midnight - CET (Central European Time) Good luck for the final spurt! :) User2776
Hi Lars!! ugh ... I know. I have been missing due to exams by a very long time! So I was under the obligation to raise my applications soon. Well, now I have enough free time, so I'll get to work on the update of my applications and discuss all your Tabajares (I've seen that there are some very good things:).) Right now I'm busy with paperwork at the university, but tonight I promise to catch up! :) User2830
Hi lars, Thx for your comments. Its always valuable for me. User3148
Hi Lars, do you use then the Loader or URLLoader class to load your request? You should use URLLoader. Stupid question, but your webspace does support PHP? Have no more ideas at the moment... User2842
Hi Lars!! Thanks for your comment :) Tonight I'll see your latest works and I comment it. See you! ;) User2830
Ah, problem fixed - super! And a really clean/elegant GUI. Do you work for Apple? User3194
Yes in a sql database.. Thanks for your comment and ill be adding that skip button in a minute.. User3186
sorry misspelled your name.. User3186
lats, i see that you just tested my app, i know its just a basic app but still would like to know you feedback. User3186
Hi Lars, of course I did use papervision as this stuff is working fantastically with great performance. User2853
Hi lars, thx for visiting my profile.ur valuable comments expecting User3148
Yep, i've tried with 'local files' and 'network only' in publish settings, but not tried do load a map in player without apikey. Sorry by missing comments and Thank you!! :) User3040
Yes you are right! A forum is really missing here! Ok I will add you on Facebook and have a look at the code. I will right you as soon as I found out something. User2858
I think that the problem is because i've created in flash, and the main app-player was created in flex, so it can create security sandbox issues. User3040
Hi. I've reactivated my facebook account: http://www.facebook.com/profile.php?id=100000363039260 I'll upload some part of my code (it's big). This community really needs a forum, like nokia "calling all inovators". User3040
Ops, i've deleted my account, do you have twitter? I can sent you a DM if we follow each other. :) User3040
Lars. How do you could show another google map inside of your application? Into to my app, new maps disappears! I'm creating the same flow as you "click details, show map", in test mode they works, but in play don't. =/ User3040
Lars!! Please !! Upload your application quickly that I was late to see it :'(.... User2830
Ok, Lars (here, 6:30pm), i'm recreating the interface deployment in html and air to try to find a workaround to sandbox issues. I hope that i find it, so i'll contribute with the possible solution. I'm waiting for a contest team response. Good job! :) User3040
Yep. I'm receiving some erros too. The app-player don't show another google maps and theese maps are fully functional on air, test and swf files on my computer. I've sent an email to contest team to provide the sources of app-player, it can help us to fix some erros, i think. User3040
Hi Lars. I've made the buttons for my app. You can download a PNG (FireWorks CS4 version, editable) at http://www.ederlima.com.br/arquivos/flash/vw-app-my-ride-ederlima-buttons.png The font type is "VolksWagen HeadLine" (Regular, Bold, 20) User3040
Hi All, best thing to have is a browser like iTune since we use our iPods. User2912
lars you're absolutely right. Great idea! ;) User2830
Hi Lars, what kind of browser and flash version do you use? I can't find any "strange" behaviour in your app. User2777
Hi Lars, we will have new driving profiles by next week. Stay tuned! User2777
Hey Lars, how is it going with your coding? :) User2776
It not my first time coding an app. But it is the first time in flash :-( User2858
Great, I'm looking forward to it! Is it your first time coding an app? User2804
Wow, that would be awesome lars! maybe you are the first one submitting an app! I am excited! User2776
Hi Annett. Yes I have some more! At the moment I am coding my first app. Hope it will be ready today. User2858
Hi Lars! Cool idea, do you have some more? User2804
More ideas of this user:
affective co-driver
Popularity: 100 %
Intelligent Navigation
Popularity: 85.71 %
right before left
Popularity: 70 %
Food Order
Popularity: 57.14 %
More applications of this user:
Alfred, the emotional co-driver
Popularity: 60 %
Gear Assistant
Popularity: 87.5 %
Number Plate Recognition
Popularity: 81.82 %
Popularity: 78.57 %