Stack Exchange
log in sign up chat discuss faq users badges

Area 51 » Technology

definition commitment beta

The Machine Learning site didn't have enough activity during the beta, and has been closed. You can download the data dump of all questions here.

52 Example Questions

active newest votes
up vote 19 down vote
What mathematics are required for implementing and understanding Artificial Neural Networks?
added by qegal Aug 25 '12 at 2:17
link
4  
I appreciate the upvotes, but we should focus on getting more questions to the 10 vote threshold instead of upvoting questions that have already reached it. – qegal Jan 14 '13 at 7:51
1  
Votes can be distributed. It would be useful if you remove your vote here and vote a question with less than 10 score(if you have voted this question). – Sai Manoj Kumar Yadlapati Mar 10 '13 at 5:08
up vote 16 down vote
How can I deal with a variable number of inputs to a feedforward neural network?
added by qegal Aug 30 '12 at 16:52
link
up vote 16 down vote
In what situations would one use a SVM over a neural network?
added by qegal Nov 10 '12 at 4:45
link
up vote 13 down vote
What is the difference between statistics and machine learning?
added by qegal Oct 27 '12 at 1:07
link
up vote 12 down vote
Is a simple naive bayesian classifier going to be accurate enough for a simple Twitter sentiment engine?
added by Steve H. Dec 9 '12 at 14:38
link
up vote 12 down vote
How do you know what algorithm is most appropriate for the data you have? Any suggestion on what should be the important steps to take?
added by zkan Jan 12 '13 at 8:27
link
up vote 12 down vote
Is there any machine learning tool/library for distributed computing?
added by zkan Jan 12 '13 at 8:46
link
up vote 12 down vote
What machine learning libraries use GPGPU to speed up computation time?
added by Benjamin Manns Jan 13 '13 at 19:23
link
up vote 12 down vote
Soft computing models - Where and how use the fuzzy models?
added by István Simon Jan 28 '13 at 16:19
link
up vote 11 down vote
How does the number of layers and number of neurons per layer affect the accuracy of feedforward backpropogation neural networks?
added by qegal Aug 30 '12 at 16:51
link
up vote 11 down vote
Soft computing models - How use the Ada-boost to create a Fuzzy system?
added by István Simon Jan 29 '13 at 8:33
link
up vote 11 down vote
Is there a decision matrix to choose a ML approach based on a problem's characteristics?
added by Sebi Feb 12 '13 at 21:57
link
up vote 10 down vote
How can i benchmark accuracy and performance of a recommender system implementation?
added by phoxis Jul 10 '12 at 8:25
link
this question is on-topic at CS.SE – Artem Kaznatcheev Jul 17 '12 at 22:24
How, it should be close to computational science, more appropriately in this forum. I was talking about the mean absolute deviation, F1 score, precision recall, etc measures, and how to use it in the case of a binary recommender and a n scale recommender. – phoxis Jul 18 '12 at 8:11
up vote 10 down vote
What are your best practices to handling user's expectations about your ML system, so they know how accurate it is and in which cases it may fail?
added by THelper Nov 14 '12 at 8:06
link
up vote 10 down vote
What type of clustering algorithm one would use for objects defined by antisymmetric non-transitive relation with each other?
added by Krzysztof Hasiński Nov 19 '12 at 13:17
link
up vote 10 down vote
Does anyone know where I can find an open source implementation of a fisher classification algorithm in Pyton 2.x?
added by Steve H. Dec 9 '12 at 15:03
link
4  
This seems like a shopping question to me. If the requested code exists, the answer will be nothing more than a link to another site (which is prone to linkrot) and if it doesn't exist, the question will remain unanswered. – Ladadadada Feb 3 '13 at 10:45
up vote 10 down vote
Is machine learning just another name for the discipline of artificial intelligence? If not, how do they relate to each other?
added by stacky-bit Dec 21 '12 at 13:38
link
2  
I know it's not a place to answer questions but this one seems more like a meta "What is this proposal about" than one to fit the actual site, so: It's a subset of Artificial Intelligence problems along with the likes of image recognition, decision-making, parsing natural language, abstraction (raw data into tokenized ideas), deduction, adaptative systems, chatterbots and many more. – SF. Jan 17 '13 at 12:40
up vote 10 down vote
Is there any recommended machine learning book?
added by zkan Jan 12 '13 at 8:19
link
The Elements of Statistical Learning by Hastie et al. is a nice one. I also enjoy Data Mining: Practical Machine Learning Tools and Techniques by Witten&Frank – Alina Mar 14 '13 at 15:29
up vote 10 down vote
How do we deal with missing values in support vector machines?
added by zkan Jan 12 '13 at 8:22
link
up vote 10 down vote
When we face an unbalanced dataset, what do we do and how do we evaluate the classification algorithm?
added by zkan Jan 12 '13 at 8:43
link
up vote 10 down vote
How can a SVMs resemble KNNs ?
added by Ojtwist Jan 13 '13 at 10:02
link
up vote 10 down vote
What is the simplest way to implement a basic Hidden Markov model from scratch?
added by slayton Jan 14 '13 at 1:08
link
up vote 10 down vote
In what real world applications are Hopfield networks used?
added by Sunny88 Jan 14 '13 at 18:09
link
Would be good to know, but is this a good fit for a Q&A? There is no single right answer. I believe something like "under what circumstances would a ... be used" would be more appropriate. – Dukeling Mar 27 '13 at 20:47
up vote 10 down vote
How to simplify algorithms generated by Genetic Programming?
added by Szilágyi Donát Jan 15 '13 at 8:00
link
The algorithms themselves aren't generated by genetic programming, are they? That's more like evolutionary programming, where the program itself can change, according to what I've heard. If it's about genetic programming it's just an algorithm with evolving parameters. Broad question. – 5th Jan 21 '13 at 13:19
In GP the algorithms are really evolved. A mutation can change a + to -, a movmin function to a movavg or can change even a big part of an algorithm to another one. GP builds a tree structure of each algorithm and modifies entire nodes, not only params. – Szilágyi Donát Jan 21 '13 at 13:55
up vote 10 down vote
Is there any work that has applied Restricted Boltzmann Machines in a regression scenario?
added by Srikanth Jan 24 '13 at 14:43
link
up vote 10 down vote
Is there a rule of thumb to select the number of layers and the numbers of neurons per layer in a Neural Network?
added by SRKX Jan 25 '13 at 11:11
link
up vote 10 down vote
What kind of skills/knowledge does someone who works with machine learning need to possess?
added by Marc DiMillo Jan 26 '13 at 8:27
link
up vote 10 down vote
Soft computing models - How use the GE to create a Fuzzy system?
added by István Simon Jan 29 '13 at 8:29
link
up vote 10 down vote
How choose the proper fitness function?
added by István Simon Jan 29 '13 at 8:45
link
up vote 10 down vote
Inspired by the Human Brain Project: How many layers of neurons does the human brain have and what is the range of the number of neurons per layer?
added by stacky-bit Jan 30 '13 at 14:29
link
up vote 10 down vote
What machine learning algorithms are used in training set based image recognition?
added by Pavel V. Jan 31 '13 at 10:07
link
up vote 10 down vote
What Java libraries exist to implement a neural network?
added by Sebi Feb 12 '13 at 21:53
link
Is it supposed to be "neural network"? – Sai Manoj Kumar Yadlapati Mar 10 '13 at 4:48
yes, changed the question – Sebi Mar 10 '13 at 19:05
Can consider Apache Mahout? mahout.apache.org – Debmalya Jash Mar 22 '13 at 3:50
Would be good to know, but is this a good fit for a Q&A? There is no single right answer. Questions like these get closed on StackOverflow very quickly. I should note that I'm not really against including these types of questions. – Dukeling Mar 27 '13 at 20:43
up vote 10 down vote
Does any library implement the Neuron By Neuron (NBN) learning algorithm for Feedforward Artificial Neural Networks ?
added by phoxis Feb 19 '13 at 16:34
link
up vote 10 down vote
In active learning, what should be considered when trying to determine appropriate stopping criteria?
added by Wesley Baugh Feb 28 '13 at 21:32
link
up vote 10 down vote
What approaches are there to learn from unlabeled data if we only have labeled instances from the positive class?
added by Wesley Baugh Feb 28 '13 at 21:45
link
up vote 10 down vote
Is it possible to do feature selection (find most informative features; discount "noisy" features) when using an online / streaming algorithm?
added by Wesley Baugh Feb 28 '13 at 21:52
link
This questions seems vague, consider specify a bit more. – RSFalcon7 Apr 8 '13 at 16:48
up vote 10 down vote
How does SVM really work?
added by Sai Manoj Kumar Yadlapati Mar 10 '13 at 5:35
link
up vote 10 down vote
What is a good Text on Probability basics required for understanding the working of Machine Learning Algorithms?
added by Sai Manoj Kumar Yadlapati Mar 10 '13 at 5:39
link
up vote 10 down vote
What are the differences between Machine Learning and Data Mining?
added by Oleksi Mar 15 '13 at 4:36
link
up vote 10 down vote
Is "Deep Learning" just another name for "Artificial Neural Networks"?
added by DaemonMaker Mar 30 '13 at 18:45
link
1 2