I've read the book Programming Collective Intelligence and found it fascinating. I'd recently heard about a challenge amazon had posted to the world to come up with a better recommendation engine for their system.
The winner apparently produced the best algorithm by limiting the amount of information that was being fed to it.
As a first rule of thumb I guess... "More information is not necessarily better when it comes to fuzzy algorithms."
I know's it's subjective, but ultimately it's a measurable thing (clicks in response to recommendations).
Since most of us are dealing with the web these days and search can be considered a form of recommendation... I suspect I'm not the only one who'd appreciate other peoples ideas on this.
In a nutshell, "What is the best way to build a recommendation ?"
You don't want to use "overall popularity" unless you have no information about the user. Instead, you want to align this user with similar users and weight accordingly.
This is exactly what Bayesian Inference does. In English, it means adjusting the overall probability you'll like something (the average rating) with ratings from other people who generally vote your way as well.
Another piece of advice, but this time ad hoc: I find that there are people where if they like something I will almost assuredly not like it. I don't know if this effect is real or imagined, but it might be fun to build in a kind of "negative effect" instead of just clumping people by similarity.
Finally there's a company specializing in exactly this called SenseArray. The owner (Ian Clarke of freenet fame) is very approachable. You can use my name if you call him up.
There is an entire research area in computer science devoted to this subject. I'd suggest reading some articles.
Agree with #Ricardo. This question is too broad, like asking "What's the best way to optimize a system?"
One common feature to nearly all existing recommendation engines is that making the final recommendation boils down to multiplying some number of matrices and vectors. For example multiply a matrix containing proximity weights between users by a vector of item ratings.
(Of course you have to be ready for most of your vectors to be super sparse!)
My answer is surely too late for #Allain but for other users finding this question through search -- send me a PM and ask a more specific question and I will be sure to respond.
(I design recommendation engines professionally.)
#Lao Tzu, I agree with you.
According to me, recommendation engines are made up of:
Context Input fed from context aware systems (logging all your data)
Logical reasoning to filter the most obvious
Expert systems that improve your subjective data over the period of time based on context inputs, and
Probabilistic reasoning to do decision-making close-to-proximity based on weighted sum of previous actions(beliefs, desires, & intentions).
P.S.
I made such recommendation engine.
Related
We are creating a website for a client that wants a website based around a survey of peoples' '10 favourite things'. There are 10 questions that each user must answer, e.g. 'What is your favourite colour', 'Who is your favourite celebrity', etc., and then the results are collated into a global Top 10 list on the home page.
The conundrum lies in both allowing the user to input anything they want, e.g. their favourite holiday destination might be 'Grandma's house', and being able to accurately count the votes accurately, e.g. User A might say their favourite celebrity is 'The Queen' and User B might says it's 'Queen of England' - we need those two answers to be counted as two votes for the same 'thing'.
If we force the user to choose from a large but predetermined list for each question, it restricts users' ability to define literally anything as their 'favourite thing'. Whereas, if we have a plain text input field and try to interpret answers after they have been submitted, it's going to be much more difficult to count votes where there are variations in names or spelling for the same answer.
Is it possible to automatically moderate their answers in real-time through some form of search phrase suggestion engine? How can we make sure that, if a plain text field is the input method, we make allowances for variations in spelling?
If anyone has any ideas as to possible solutions to this functionality, perhaps a piece of software, a plugin, an API, anything, then please do let us know.
Thank you and please just ask for any clarification.
If you want to automate counting "The Queen" and "The Queen of England", you're in for work that might be more complex than it's worth for a "fun little survey". If the volume is light enough, consider just manually counting the results. Just to give you a feeling, what if someone enters "The Queen of Sweden" or "Queen Letifah Concerts"?
If you really want to go down that route, look into Natural Language Processing (NLP). Specifically, the field of categorization.
For a general introduction to NLP, I recommend the relevant Wikipedia article
http://en.wikipedia.org/wiki/Natural_language_processing
RapidMiner is an open source NLP solution that would be worth looking into.
As Eric J said, this is getting into cutting edge NLP applications. These are fields of study that are very important for AI/automation researchers and computer science in general, but are still very fledgeling. There are a number of programs and algorithms you can use, the drawbacks and benefits of which very widely. RapidMiner is good, WordNet is widely used in medical applications and should be relatively easy to adjust to your own corpus, and there are more advanced methods like latent Dirichlet allocation. Here are a few resources you should start with (in addition to the Wikipedia article provided above)
http://www.semanticsearchart.com/index.html
http://www.mitpressjournals.org/loi/coli
http://marimba.d.umn.edu/ (try the SenseClusters calculator)
http://wordnet.princeton.edu/
The best to classify short answers is k-means clustering. You need to apply stemming. Then you need to convert words into indexes using elementary dictionary. You can use EverGroingDictionary.cs from sematicsearchart.com. After throwing phrase to a dictionary it will be converted to sequence of numbers or vector. Introduce measure of proximity as number of coincidences in words and apply k-means, which is lightning fast algorithm. k-means will organize all answers into groups. Most frequent words in each group will be a signature of the group. Your whole program in C++ or C# or Java must be less than 1000 lines.
Simple item-to-item recommendation systems are well-known and frequently implemented. An example is the Slope One algorithm. This is fine if the user hasn't rated many items yet, but once they have, I want to offer more finely-grained recommendations. Let's take a music recommendation system as an example, since they are quite popular. If a user is viewing a piece by Mozart, a suggestion for another Mozart piece or Beethoven might be given. But if the user has made many ratings on classical music, we might be able to make a correlation between the items and see that the user dislikes vocals or certain instruments. I'm assuming this would be a two-part process, first part is to find correlations between each users' ratings, the second would be to build the recommendation matrix from these extra data. So the question is, are they any open-source implementations or papers that can be used for each of these steps?
Taste may have something useful. It's moved to the Mahout project:
http://taste.sourceforge.net/
In general, the idea is that given a user's past preferences, you want to predict what they'll select next and recommend it. You build a machine-learning model in which the inputs are what a user has picked in the past and the attributes of each pick. The output is the item(s) they'll pick. You create training data by holding back some of their choices, and using their history to predict the data you held back.
Lots of different machine learning models you can use. Decision trees are common.
One answer is that any recommender system ought to have some of the properties you describe. Initially, recommendations aren't so good and are all over the place. As it learns tastes, the recommendations will come from the area the user likes.
But, the collaborative filtering process you describe is fundamentally not trying to solve the problem you are trying to solve. It is based on user ratings, and two songs aren't rated similarly because they are similar songs -- they're rated similarly just because similar people like them.
What you really need is to define your notion of song-song similarity. Is it based on how the song sounds? the composer? Because it sounds like the notion is not based on ratings, actually. That is 80% of the problem you are trying to solve.
I think the question you are really answering is, what items are most similar to a given item? Given your item similarity, that's an easier problem than recommendation.
Mahout can help with all of these things, except song-song similarity based on its audio -- or at least provide a start and framework for your solution.
There are two techniques that I can think of:
Train a feed-forward artificial neural net using Backpropagation or one of it's successors (e.g. Resilient Propagation).
Use version space learning. This starts with the most general and the most specific hypotheses about what the user likes and narrows them down when new examples are integrated. You can use a hierarchy of terms to describe concepts.
Common characteristics of these methods are:
You need a different function for
each user. This pretty much rules
out efficient database queries when
searching for recommendations.
The function can be updated on the fly
when the user votes for an item.
The dimensions along which you classify
the input data (e.g. has vocals, beats
per minute, musical scales,
whatever) are very critical to the
quality of the classification.
Please note that these suggestions come from university courses in knowledge based systems and artificial neural nets, not from practical experience.
My friend works for a non-profit organization working to stop the illegal exploitation of minors over sites such as craigslist.org, which is one of the more popular mediums. The question is whether or not it is possible, now or in the near future, to develop an algorithm to analyze a photo of a person and return a prediction of their relative age.
It sounds like a mammoth task. My only thought was some sort of Bayesian probability system. I know even people often have trouble judging someone's age but Bayesian spam filters are advertised as being "10 times as accurate as a human" so maybe it's possible?
I am pretty inexperienced though. I would appreciate it if someone else could suggest whether or not this is feasible and if so how and when?
EDIT: Thank you everyone for the responses. Smoore that study was very helpful but I think Hal's solution is the most practical for the time being.
Here's a possible (left-field) solution. Perhaps, you could tie it into some type of a captcha solution for the site itself. Prompt new users with images of other new users with the question: "Is this person over 18?". It's true that a 50% success rate is not a very effective captcha system, but it's a start.
Coupled with some other checks or repetitive checks and it could work. You could display the image to a number of new users, and base the result on a certain threshold. If, 8 out of 10 people flagged a certain image as not a minor, than it's probably pretty safe they are of age.
But, this whole system can be circumvented by simply uploading someone else's image so I'm not sure how effective any of this really is. :)
I expect it would be pretty hard to get right. Consider this set of photos where the same model is made up to look very different ages.
There are algorithm to reliably determine the attractiveness of a face. See acm.org and uni-regensburg.de. It wouldn't be too much of a stretch to imagine an algorithm which could predict age.
Characteristics such as smoothness would probably have a strong correlation with age. It would probably take a great deal of effort to be more reliable than your average carney though.
I think you would need some input from a forensic anthropoligist ( or at least an anatomist).
Differnet parts of the body grow at different rates so it might be possible to do something like size of head vs. shoulder width, arm length vs. body width.
Unfortunately it sounds like he is trying to differentiate between say a 14 year olds and 18 year olds. Which is only a four year difference, variations in genetic makeup and nutitrition would probaly give any system an accuracy of +/- 20% which would equate to three years for this age group.
On the other hand if you had a large sample of photos then you could account for the variance statisticaly and get a pretty good idea whether a site was likely to be exploiting minors systematicaly.
The direct answer to your question is that no, no such algorithm will exist in the near future, and is probably impossible to achieve with any accuracy without strong AI.
That said, a practical solution to your problem is probably the amazon mechanical turk:
http://mturk.com
There, you can pay a small fee to have real people complete a task for you. I'd probably set your task up so that you paid $0.02 to have a person estimate the age of maybe 5 faces at a time. You could double or triple check your results with other workers, particularly for those faces who seemed close to your age limit. This is probably your only practical solution other than hiring minimum wage interns to manually review all submissions.
Use mechanical turk
In this study they tried it by analysing facial geometry and wrinkle features. Problem is this would be affected by shot angle, lighting, etc.
In some theoretical sense it is probably possible. For all practical purposes though, it is currently impossible.
Mammoth is an understatement I think. "Giant glacier" or "moon" might be more appropriate.
This isn't to say it wouldn't be worth looking into but I have a feeling you'd be in for a lot of man hours before you came up with something remotely useful.
I don't think it's something that a computer could do with any degree of accuracy. It's even really hard for people to do. I mean, have you been the the liquor store lately, they are supposed to ask for ID from anybody who looks under 25 (drinking age is 19 here). Apparently some 40 year olds don't look old enough. Telling somebody's age just by looking at them is a very hard thing to do. Especially when you get into to erotic picture arena, where they are trying to make models seem younger than they really are.
I think you will also have difficulties with different composited pictures. For instance angles on a face, different lighting, as well as context and probably most of all... image quality/resolution. It's a lot easier to work with a 800x600 pic then it is to work with a 320x240. The algorithm is only as good as the subject.
I cannot see this approach (a software solution to measuring age) being very effective. I like the idea of users flagging images - a human being can discern age many times more effectively then any algorithm.
Practical approach aside, I'd advice against trying to develop anything in that direction for now.
Few reasons:
1. guessing someone's age is not a grateful task
2. "biological" age and "calendar" age of people vary greatly - I know people who are 30 and are still asked for an ID when buying liquor, and some who are barely 18 and already look over 30
3. some people's looks don't change over time - they just have that kind of looks
4. nowadays, everyone's working to look as young as they can - so basically, you've got the whole industry working against you :(
Anyways, to cut long story short, I don't think it's feasible for now.
A neural net is a reasonable approach, you would need a training set of pictures of people with known ages and a bit of image processing to remove hats etc.
edit: Question changed?
You might be ale to classify someone as 20-30 or 40-50 on a CCTV but you aren't going to be ale to tell if a model is 17 or 18 in a posed photo.
Just like nearly all advanced tasks in image classification this topic is still in research. Judging from this paper it is possible to do it but non-trivial, also you have to have a lot of (manually) annotated training data. Without any knowledge of this field and no experience in image processing this task is going to take you several months.
Develop a classification algorithm that bases a heuristic on many values of the pictures, amount of pixels that are dark within the face area (possibly wrinkles), and the color of the hair. These values should fall within a general area of any profile-esque picture, if you want to be fancy, carry weights with these values and develop a type of game tree that would be able to search hundreds of thousands of images quickly, finding where this image "falls" in the tree within an age-specific set of values.
Some Japanese cigarette vending machines do this. Not terribly well by all accounts, but then it probably doesn't matter since, as Hal mentioned, the easiest hack is just to use someone else's image...
Impossible is nothing, Only amount of efforts changes :
I think it would be near impossible if you target one particular feature of face.
you have to consider multiple factor, So decision will be lying in a matrix and you have to feed multiple things and you will get your answer i would enlist some feature :
1) Beard (Detect face , Now detect beard on face , Help full in distinguish male/female
/childern )
2) Hair
3) Wrinkles
4) Size of face
5) Ration between height and breadth of face
It would be a tough assignment but algorithm can be developed.
As of now, this is possible with 90% accuracy. Yes. please refer the following link..
http://www.omron.com/r_d/coretech/vision/okao.html
I'm working on a web application which will be used for classifying photos of automobiles. The users will be presented with photos of various vehicles, and will be asked to answer a series of questions about what they see. The results will be recorded to a database, averaged, and displayed.
I'm looking for algorithms to help me identify users which frequently don't vote with the group, indicating that they're probably either not paying attention to the photos, or that they're lying about what they see. I then want to exclude these users, and recalculate the results, such that I can say, with a known amount of confidence, that this particular photo shows a vehicle that is this and that.
This question goes out to all you computer science guys, where to find such algorithms or to give myself the theoretical background to design such algorithms. I'm assuming I'm going to have to learn some probability and statics, maybe some data mining. Some book recommendations would be great. Thanks!
P.S. These are multiple choice questions.
All of these are good suggestions. Thank you! I wish there was a way on stack overflow to select multiple correct answers so more of you could be acknowledged for your contributions!!
Read The Elements of Statistical Learning, it is a great compendium on data mining.
You can be interested especially in unsupervised algorithms, for example clustering. Assuming that most people do not lie, the biggest cluster is right and the rest is wrong. Mark people accordingly, then apply some bayesian statistics and you'll be done.
Of course, most data mining technologies are pretty experimentative, so don't count on that they will be always right... or even in most cases.
I believe what you described is solved using outlier/anomaly detection.
A number of techniques exist:
statistical-based methods
distance-based methods
model-based methods
I suggest you take a look at these slides from the excellent book Introduction to Data Mining
If you know what answers you are expecting why do you ask people to vote? By excluding some values you basically turn the vote in something that you like. Automobiles make different impression to different individuals. If 100 ppl loved a car then when someone comes and says that he/she doesn't like it, you exclude the vote?
But anyway, considering that you still want to do this, first of all you will need a large set o data from "trusted" voters. This will give you an idea of "good" answer and from this point you can choose the exclude threshold.
Without an initial set of data you cannot apply any algorithm because you will get false results. Consider just one vote of 100 from on a scale from 0 to 100. The second vote is "1" The you will exclude this vote because is too far away from the average.
I think a pretty simple algorithm could accomplish this for you. You could try and get fancier by calculating the standard deviations and such, but I wouldn't bother.
Here's a simple approach that should be sufficient:
For each of your users, calculate the number of questions they answered and the number of times they selected the most popular answer for the question. The users which have the lowest ratio of picking the popular answer versus total answers you can guess are providing bogus data.
You probably would not want to throw out the data from users where they've only answered a small number of questions because they likely have just disagreed on a few versus putting in bogus data.
What kind of questions are they (Yes/No, or 1 to 10?).
You may be able to get away with not discarding anything by using a mean instead of an average. With averages if there are extreme outliers in the response it could affect the average, but if you use median you may get a better answer. So for example if you had 5 answers, order them and pick the middle one.
I think what you are saying is that you are concerned that certain people are "outliers", and they are adding noise to your data, making the categorizations less reliable. So, if you have a Chevy Camaro, and most people say it is either a pony car, a muscle car, or a sports car, but you have some goofball who says it's a family sedan, you would want to minimize the impact of his vote.
One thing you could do is provide a Stack Overflow-like reputation score for users:
The more a user is "in agreement" with other users, the better his or her score would be. For a given user (User X), this could be determined by a simple calculation of what percentage of users who responded to a question chose the same category as User X, then averaging this value over all questions answered.
You may want to to multiply this value by the total number of question answered to encourage people to answer as many questions as possible. (Note: if you choose to do this, it would be equivalent to just summing the percentage agreement scores rather than averaging them.)
You could present the final reputation score to users, making sure to explain that they will be rewarded for how well their responses agree with those of other users. This will encourage people to answer more questions but also to take care in their answers.
Finally, you could calculate a certainty score for a given categorization by adding up the total reputation score of all people who chose a given category.
Some of these ideas may need some refinement, especially since I don't know your exact situation. Certainly, if people can see what other people chose before they vote, it would be way too easy to game the system.
If you were to collect votes like "on a scale from 1 to 10, how would you rate this car", you could probably use simple average and standard deviation: the smaller the standard deviation, the more unanimous the general consensus is among your voters, and you can flag users who are e.g. 3 standard devs from the average.
For multiple choice, you need to be more careful. Simply discarding all but the most-voted option will do nothing but disgruntle the voters. You need to establish a measure of how significant the winner is w.r.t. the other options, e.g. flag users who voted for options with less than 1/3 of the winning options count.
Note that I wrote "flag users", not discard votes. If you discard votes, you can't tell how confident you are about the result ("91% voted this to be a Ford Mustang"). If a user has more than a certain percentage of his votes flagged - well, that's up to you.
Your trickiest problem, however, will probably be to collect sufficient votes. Depending on how easy the multiple choice problem is, you probably need several times the number of options as votes, per photo. Otherwise the statistics are meaningless.
We are embarking on some R&D for a staff rostering system, and I know that there are some suggested algorithms such as the memetic algorithm etc., but I cannot find any additional information on the web.
Does anyone know any research journals, or pseudocode out there which better explains these algorithms?
Thanks,
Devan
Here is a useful document:
Memetic Algorithms for Nurse Rostering (pdf)
It contains a little bit of theory and pseudo-code.
Scheduling problem is NP-hard and usually being solved using genetic algorithms (GA).
You can start learning GA from Wikipedia article
You may also want to look at a technique called "simulated annealing". Like genetic algorithms, this uses an evaluation function to determine the quality of candidate solutions - but the generating of the candidates tends to be simpler. Each type of algorithm gives better results in certain circumstances - from a brief Google survey it feels like genetic has the edge, but annealing will be quicker to implement.
Here is a comparison paper (for a different domain, not scheduling):
http://www.ee.utulsa.edu/~tmanikas/Pubs/gasa-TR-96-101.pdf
We have used simulated annealing in a large scheduling application and it did work well.
To be honest, if the volume of staff is less than about 40, I would recommend giving a visual representation of the roster and letting the user finalise the schedule. Perhaps you would use an algorithm to produce a candidate schedule to start with, and then let the user play with it. You could still use the evaluation function to check the user's work and give feedback on how good their solution is.
There are many many many issues to consider when setting up a roster schedule, so aku's tip about genetic algorithms is the best one.
You need a good evaluation function to determine the quality of the roster for such an algorithm, and you can, and should, consider things like the following (but not limited to):
have you solved the workload problem with this roster? (ie. do you have enough people at work at all times?)
if not, can you live with the consequences? (for hospitals, you might have to postpone lunch 15 min one day in order to have enough people available for it, or just drag it slightly out in time)
is the roster a good one, considering things like shift stability for each person, their days off, whether or not they get weekends off with some regularity
is the roster legal? taking things like local regulations into account, that regulate things like how much time must pass between one shift and another (downtime), how much can each person work inside a given interval (day, week, month)
I read a rostering algo paper by these guys a while back.
Or by using OR ;)