I am looking sites\blogs where people explain how they solving perfomance and other problems with big data. I know some resources about scalable web applications and sites(like twitter, facebook). it's ok but i am looking concrete algorithms mainly for data mining.
Actually a lot of the things done on big data do no qualify as "data mining".
At most they apply previously learned rules to uniform big data, mostly in predicting consumer interest to serve them ads. But that mostly boils down to "has interest in sports" type of decisions. Quite a huge error rate is acceptable here, as there is next to no cost to serving someone a sports ad who is not interested in sports. The web is full of stories where Google put people into the wrong consumer segments. Often even predicting the gender incorrectly.
Whenever you see "big data", take that with a big grain of salt. It mostly is bragging and buzzword bingo. The challenge with big data still lies in actually getting it done, not (yet) with getting it done right.
A good example is this article: http://www.technologyreview.com/web/39487/
Yahoo predicted (using Twitter "big data" and pushing this article that claims they are much better than exit polls etc. brag brag brag) that "romney has a 90% chance of winning south carolina". In reality, Romney got 28%, while Gingrich got 40%.
Or try out some of the "sentiment analysis" type of tools. They will tell you that a twitter post containing "puppies" is positive, and containing "cockroaches" is negative. And that is about the quality they get with "sentiment analysis" these days. Again, they are so much focused on just getting anything out of the data, they are not yet at the point of actually analyzing (or even validating) the results. Sorry. I bet I'll get some downvotes for being this critical, but this is what is happening everyday. See the yahoo example. They apparently were able to process their "big data", but their results surely aren't ready for prime time, they still need to work on that.
And again, for some situations such as ad targeting, error rates can be quite high. Anything better than random, is, well, better than random! Which means more money than if you would just serve random ads. So it isn't worthless; just maybe not comparable to non-big-data approaches.
Related
We have an application with 10 millions lines of code in 4GL(Progress) and a database also OpenEdge with 300 Tables. My Boss says we should migrate it to a new Programming language and a new Database Management system.
My questions are:
Do you think we should migrate it? Do you think Progress has a "future"?
If we should migrate it, how, are there any tools? Or should we begin with programming from scratch?
Thank you for the help.
Ablo
Unless your boss has access to an unlimited budget, endless user patience and a thirst for frustration and agony you should not waste any time thinking about rewrites.
http://www.joelonsoftware.com/articles/fog0000000069.html
Yes, Progress has a future. They probably will never be as sexy an option as Microsoft or Oracle or whatever the cool kids are using this week. But they have been around for 30 years and they will still be here when you and your boss retire.
There are those who will rain down scorn on Progress because it isn't X or it doesn't have Y. Maybe they can rewrite your 10 million lines of code next weekend and prove just how right they are. I would not, however, pay them for those efforts until after the user acceptance tests are passed and the implementation is completed.
A couple of years later (the original post being from 2014 and the answers being from 2014 to 2015) :
The post, which has gotten the most votes is argumenting basically two fold :
a. Progress (Openedge) has been around for a long time and is not going anywhere soon
b. Unless your boss has access to an unlimited budget, endless user patience and a thirst for frustration and agony you should not waste any time thinking about rewrites: http://www.joelonsoftware.com/articles/fog0000000069.html
With regard to a:
Yes, the Progress OpenEdge Stack is still around. But from my experience the difficulty to find experienced and skilled Openedge has gotten even more difficult.
But also an important factor here, which i think has evolved to much greater importance, since this discussion started:
The available Open Source Stacks for application development have gotten by factors better, both in terms of out-of-box functionality and quality and have decisively moved in direction of RAD.
I am thinking for instance of Spring Boot, but not only, see https://stackshare.io/spring-boot/alternatives. In the Java realm Spring Boot is certainly unique. Also for the development of rich Webui's many very valid options have emerged, which certainly are addressing RAD requirements, just some "arbitrary" examples https://vaadin.com for Java, but also https://www.polymer-project.org for Javascript, which are interestingly converging both with https://vaadin.com/flow.
Many of the available stacks are still evolving strongly, but all have making life easier for the developer as strong driver. Also in terms of architectures you will find a convergence of many of this stacks with regard basic building blocks and principles: Separation of Interfaces from Implementation, REST API's for remote communication, Object Relational Mapping Technologies, NoSql / Json approaches etc etc.
So yes the Open Source Stack are getting very efficient in terms of Development. And what must also be mentioned, that the scope of these stacks do not stop with development: Deployment, Operational Aspects and naturally also Testing are a strong ,which in the end also make the developers life easier.
Generally one can say the a well choosen Mix and Match of Open Source Stacks have a very strong value proposition, also on the background of RAD requirements, which a proprietary Stack, will have in the long run difficulty to match - at least from my point of view.
With regard to b:
Interestingly enough i was just recently with a customer, who is looking to do exactly this: rewrite their application. The irony: they are migrating from Progress to Progress OpenEdge, with several additional Open Edge compliant Tools. The reason two fold: Their code is getting very difficult to maintain and would refactoring in order to address requirements coming from Web Frontends. Also interesting, they are not finding enough qualified developers.
Basically: Code is sound and lives , when it can be refactored and when it can evolve with new requirements. Unfortunately there many examples - at least from my experience - to contrary.
Additionally End-of-Lifecyle of Software can force a company, to "rewrite" at least layers of their software. And this doesn't necessarily have to bad and impossible. I worked on a Project, which migrated over 300 Oracle Forms forms to a Java based UI within less then two years. This migration from a 2 tier to a 3 tier architecture actually positioned the company to evolve their architecture to address the needs of Web Ui's. So actually in the end this "rewrite" and a strong return of value also from the business perspective.
So to cut a (very;-)) long story short:
One way or another, it is easy to go wrong with generalizations.
You need not begin programming from scratch. There is help available online and yes, you can contact Progress Technical Support if you find difficulties. Generally, ABL code from previous version should work with only little changes. Here are few things that you need to do in order to migrate your application:
Backup databases
Backup source code and .r files
Truncate DB bi files
Convert your databases
Recompile ABL code and test
http://knowledgebase.progress.com articles will help you in this. If you are migrating from some older versions like 9, you can find a good set of new features. You can try them but only after you are done with your conversion.
If you are migrating from 32-bit to 64-bit and if you are using 32-bit libraries, you need to replace them with 64-bit
The first question I'd come back with is 'why'? If the application is not measuring up that's one thing, and the question needs to be looked at from that perspective.
If the perception is that Progress is somehow a "lesser" application development and operating environment, and the desire is only to move to a different development and operating environment - you'll end up with a lot of resources in time, effort, and money invested - not to mention the opportunity cost - and for what? To run on a different database platform? Will migrating result in a lower TCO? Faster development turn-around time? Quicker time to market? What's expected advantage in moving from Progress, and how long will it take to recover the migration cost - if ever?
Somewhere out there is a company who had similar thoughts and tried to move off of Progress and the ABL. The effort failed to meet their target performance and functionality metrics, so they eventually gave up on the migration, threw in the towel, and stayed with Progress - after spending $25M on the project.
Can your company afford that kind of risk / reward ratio?
Progress (Openedge) has been around for a long time and is not going anywhere soon. And rewriting 10 Million lines of code in any language just to use the current flavor of the month would never be worth it unless your current application is not doing what you need. Even then bringing it up to current needs would normally be a better solution.
If you need to migrate your current application to the latest version of Openedge (Progress) you would normally just make a copy of your database(s) and convert it/them to the new version of Openedge and compile your your code against the new databases and shake the bugs out. You may have some keyword issues, but this is usually pretty minor.
If you need help with programming I would suggest contacting Progress Software and attending the yearly trade show or going to https://community.progress.com/ and asking/looking for local user groups. The local user groups would be a stellar place to find local programming talent.
Hope this helps.....
I am really passionate about the machine learning,data mining and computer vision fields and I was thinking at taking things a little bit further.
I was thinking at buying a LEGO Mindstorms NXT 2.0 robot for trying to experiment machine learning/computer vision and robotics algorithms in order to try to understand better several existing concepts.
Would you encourage me into doing so? Do you recommend any other alternative for a practical approach in understanding these fields which is acceptably expensive like(nearly 200 - 250 pounds) ? Are there any mini robots which I can buy and experiment stuff with?
If your interests are machine learning, data mining and computer vision then I'd say a Lego mindstorms is not the best option for you. Not unless you are also interested in robotics/electronics.
Do do interesting machine learning you only need a computer and a problem to solve. Think ai-contest or mlcomp or similar.
Do do interesting data mining you need a computer, a lot of data and a question to answer. If you have an internet connection the amount of data you can get at is only limited by your bandwidth. Think netflix prize, try your hand at collecting and interpreting data from wherever. If you are learning, this is a nice place to start.
As for computer vision: All you need is a computer and images. Depending on the type of problem you find interesting you could do some processing of random webcam images, take all you holiday photo's and try to detect where all your travel companions are in them. If you have a webcam your options are endless.
Lego mindstorms allows you to combine machine learning and computer vision. I'm not sure where the datamining would come in, and you will spend (waste?) time on the robotics/electronics side of things, which you don't list as one of your passions.
Well, I would take a look at the irobot create... well within your budget, and very robust.
Depending on your age, you may not want to be seen with a "lego robot" if you are out of college :-)
Anyway, I buy the creates in batches for my lab. You can link to them with a hard cable(cheap) or put a blue tooth interface on it.
But a webcam on that puppy, hook it up to a multicore machine and you have an awesome working robot for the things you want to explore.
Also, the old roombas had a ttl level serial port (if that did not make sense to you , then skip it). I don't know about the new ones. So, it was possible to control any roomba vacuum from a laptop.
The Number One rule, and I cannot emphasize this enough: Have a reliable platform for experimentation. If you hand build something, just for basic functionality, you will spend all your time on minor issues and not get to the fun stuff.
Anyway. best of luck.
I'm on a project that among other video related tasks should eventually be capable of extracting the audio of a video and apply some kind of speech recognition to it and get a transcribed text of what's said on the video. Ideally it should output some kind of subtitle format so that the text is linked to a certain point on the video.
I was thinking of using the Microsoft Speech API (aka SAPI). But from what I could see it is rather difficult to use. The very few examples that I found for speech recognition (most are for Text-To-Speech which mush easier) didn't perform very well (they don't recognize a thing). For example this one: http://msdn.microsoft.com/en-us/library/ms717071%28v=vs.85%29.aspx
Some examples use something called grammar files that are supposed to define the words that the recognizer is waiting for but since I haven't trained the Windows Speech Recognition thoroughly I think that might be adulterating the results.
So my question is... what's the best tool for something like this? Could you provide both paid and free options? Well the best "free" (as it comes with Windows) option I believe it's SAPI, all the rest should be paid but if they are really good it might be worth it. Also if you have any good tutorials for using SAPI (or other API) on a context similar to this it would be great.
On the whole this is a big ask!
The issue with any speech recognition system is that it functions best after training. It needs context (what words to expect) and some kind of audio benchmark (what does each voice sound like). This might be possible in some cases, such as a TV series if you wanted to churn through hours of speech -separated for each character- to train it. There's a lot of work there though. For something like a film there's probably no hope of training a recogniser unless you can get hold of the actors.
Most film and TV production companies just hire media companies to transcribe the subtitles based on either direct transcription using a human operator, or converting the script. The fact that they still need humans in the loop for these huge operations suggests that automated systems just aren't up to it yet.
In video you have a plethora of things that make you life difficult, pretty much spanning huge swathes of current speech technology research:
-> Multiple speakers -> "Speaker Identification" (can you tell characters apart? Also, subtitles normally have different coloured text for different speakers)
-> Multiple simultaneous speakers -> The "cocktail party problem" - can you separate the two voice components and transcribe both?
-> Background noise -> Can you pick the speech out from any soundtrack/foley/exploding helicopters.
The speech algorithm will need to be extremely robust as different characters can have different gender/accents/emotion. From what I understand of the current state of recognition you might be able to get a single speaker after some training, but asking a single program to nail all of them might be tough!
--
There is no "subtitle" format that I'm aware of. I would suggest saving an image of the text using a font like Tiresias Screenfont that's specifically designed for legibility in these circumstances, and use a lookup table to cross-reference images against video timecode (remembering NTSC/PAL/Cinema use different timing formats).
--
There's a bunch of proprietary speech recognition systems out there. If you want the best you'll probably want to license a solution off one of the big boys like Nuance. If you want to keep things free the universities of RWTH and CMU have put some solutions together. I have no idea how good they are or how well they might be suited to the problem.
--
The only solution I can think of similar to what you're aiming at is the subtitling you can get on news channels here in the UK "Live Closed Captioning". Since it's live, I assume they use some kind of speech recognition system trained to the reader (although it might not be trained, I'm not sure). It's got better over the past few years, but on the whole it's still pretty poor. The biggest thing it seems to struggle with is speed. Dialogue is normally really fast, so live subtitling has the extra issue of getting everything done in time. Live closed captions quite frequently get left behind and have to miss a lot of content out to catch up.
Whether you have to deal with this depends on whether you'll be subtitling "live" video or if you can pre-process it. To deal with all the additional complications above I assume you'll need to pre-process it.
--
As much as I hate citing the big W there's a goldmine of useful links here!
Good luck :)
This falls into the category of dictation, which is a very large vocabulary task. Products like Dragon Naturally Speaking are amazingly good and that has a SAPI interface for developers. But it's not so simple of a problem.
Normally a dictation product is meant to be single speaker and the best products adapt automatically to that speaker, thereby improving the underlying acoustic model. They also have sophisticated language modeling which serves to constrain the problem at any given moment by limiting what is known as the perplexity of the vocabulary. That's a fancy way of saying the system is figuring out what you're talking about and therefore what types of words and phrases are likely or not likely to come next.
It would be interesting though to apply a really good dictation system to your recordings and see how well it does. My suggestion for a paid system would be to get Dragon Naturally Speaking from Nuance and get the developer API. I believe that provides a SAPI interface, which has the benefit of allowing you to swap in the Microsoft speech or any other ASR engine that supports SAPI. IBM would be another vendor to look at but I don't think you will do much better than Dragon.
But it won't work well! After all the work of integrating the ASR engine, what you will probably find is that you get a pretty high error rate (maybe half). That would be due to a few major challenges in this task:
1) multiple speakers, which will degrade the acoustic model and adaptation.
2) background music and sound effects.
3) mixed speech - people talking over each other.
4) lack of a good language model for the task.
For 1) if you had a way of separating each actor on a separate track that would be ideal. But there's no reliable way of separating speakers automatically in a way that would be good enough for a speech recognizer. If each speaker were at a distinctly different pitch, you could try pitch detection (some free software out there for that) and separate based on that, but this is a sophisticated and error prone task.) The best thing would be hand editing the speakers apart, but you might as well just manually transcribe the speech at that point! If you could get the actors on separate tracks, you would need to run the ASR using different user profiles.
For music (2) you'd either have to hope for the best or try to filter it out. Speech is more bandlimited than music so you could try a bandpass filter that attenuates everything except the voice band. You would want to experiment with the cutoffs but I would guess 100Hz to 2-3KHz would keep the speech intelligible.
For (3), there's no solution. The ASR engine should return confidence scores so at best I would say if you can tag low scores, you could then go back and manually transcribe those bits of speech.
(4) is a sophisticated task for a speech scientist. Your best bet would be to search for an existing language model made for the topic of the movie. Talk to Nuance or IBM, actually. Maybe they could point you in the right direction.
Hope this helps.
I have recently become interested in the field(s) of data mining and machine learning. The idea of going through huge datasets and trying to correlate hidden patterns and trends is fascinating. So far I have done the following
Used Weka to load simple data sets and generate decision trees
Continously read books, wiki's, blogs and SO on the same
Started playing around SQL Server DM and Python API's
Have an idea on options of freely available data sets on the web(freedb, UN etc)
What is hindering me is the minute I try to go beyond classification/associsciation and into priori/apriori algorithms I am stuck because understanding mathematical equations and logic is not(to put it modestly) one of my strong points.
So my question would be are there anybody in the Data mining field(in the role of product owner or builder) who are not naturally mathematicians? If so, how would you approach in undestanding the field since free tools like Weka and Rapid-miner both expects some mathematical/statistical background?
P.S: Excuse me if I made some mistake in the query like mixing Data mining and analytics when they are separate as I am still getting my feet wet. I hope my core question is clear.
Well, being able to do some analysis of what the data mining models are showing is absolutely vital. However, these days all of the math and statistics are taken care of by the data mining models. You don't need to understand the math behind them (although it helps).
For example, you can look through the SQL Server Analysis Services Data Mining Algorithms and see that even the technical reference is how to use these implementations, not how to recreate them.
If you can understand the business cases and you can understand what the data mining is telling you, there's really no need to delve into the math behind it.
As for some of the free tools, I've never used them, so I can't speak to them. However, I'm a big fan of SSAS and those data mining models, which don't require an extensive mathematical background.
As Eric says, and as far as you only intend to use the existing algorithms and APIs and make sense from them, I don't see problems with the required math/statistics skill set (anyway, you'll need some previous basic knowledge/level).
Now, if you intend to do research or if you want to improve or modify existing algorithms, or why not, create your own algorithms, then math and statistics is a MUST. I just started doing some research in this area, and I'm still trying to fill my skills gap =)
I can find the technical explanation of what data mining is in a book or on Wikipedia, but I'm wondering what sort of development does it exactly involve? Is it more about using tools or more about writing tools? Is it really any much different from other domains when it comes to R&D?
Data Mining is the process of discovering interesting patterns in large amounts of data. It is not querying data, which is just what user Treb describes (sorry Treb).
To understand DM from a developer's perspective, you should read the book Programming Collective Intelligence by Toby Segaran.
In my experience (I'm a former data miner :-)), it's a mixture of using tools and writing tools. A lot of the time, the tools you need to analyse the particular data set don't exist, so you have to write them yourself first. It can be very interesting but you often need quite a different approach to the sort of programming I do now (embedded wireless), for example.
You really ought to change the accepted answer on this question so it doesn't mislead those who come across it.
Saying that querying a database IS data mining because "[h]ow would you discover any pattern in your data without querying first?" is like saying opening your car door is driving because "how else would you be able to drive somewhere without opening the car door first."
You can read your data out of a text file if you want. My first data mining assignment used data sets from the UCI repository and those are almost all text files.
If you want to learn about data mining start by looking up clustering and classification. Learn about decision trees and rule based classification. Then look at k-nearest-neighbor and k-means. After that if you really want to see what data mining is all about look at Chameleon, DBScan, and Support Vector Machines. Don't necessarily learn the minutiae of the last three (they're pretty complex and math heavy) but understanding the abstract idea of what happens will tell you all you need to know in order to use the many tools and libraries that are available for each strategy.
These are only the algorithms that popped into my head just now. There are so many others that I don't recall or don't even know yet.
Data mining is about searching large quantities of data for hidden patterns. Web 2.0 example: News corp uses its site myspace.com as a large data mine to determine what movies and products to promote. They write software to identify trends in the data that it's users post to the site. News corp does this to gather information useful for advertising campaigns and market predictions. It's different from other domains of R&D in that from a data givers perspective its passive. Rather than going out on the street and asking people in person what movies they are likely to see this summer and other such questions, the data mining tools sort out these things by analyzing data given by users voluntarily.
Wikipedia actually does have a pretty good article on it:
- http://en.wikipedia.org/wiki/Data_mining
Data Mining as I say is finding patterns or trends from given data. A developer perspective might be in applications like Anti Money Laundring... Where given a pattern you will search data for that given pattern. One other use is in Projection Softwares... where you project a result or outcome in future against a heuristic by studying recognizing the current trend from data.
I think it's more about using off the shelf tools rather than developing your own. An academic example of that kind of tools might be WEKA. Of course, you still have to know what algorithms use, how to preprocess data (very important this part), etc.
In R&D I don't have much idea, but it should be like almost everything: maths, statistics, more maths...
On the development level, data mining is just another database application, but with a huge amount of data.
The mining itself is done by running specific queries on the database. It's in the creation of the queries where the important work is done. They of course depend on the data model, and on the hypotheses, what sort of trends the customer expects to find.
Therefore, the fine tuning of the queries usually can't be done in development, but only once the system is live and you have live data. Then the user can test his hypotheses and adapt the queries to show him the trends he is looking for.
So from a dev point of view, data maining is about
Managing large sets of data in your client (one query may return 100.000 rows of data)
Providing the user (who may know nothing about SQL or relational databases in general) with an effective way to modify his queries and view the results.