It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.
Closed 10 years ago.
I have a list of cities like London, Paris, Roma etc.
I want to return which cities are interesting based on some features. I asked people to judge whether they find those cities interesting or not. I show each city with the average of people's judgement of how interesting it is.
Is this supervised or unsupervised learning?
This is supervised learning.
Supervised learning is when your training set consists of both the input data and the desired output data. In this case, the input data is the cities and the output data is the ratings that the people have given it.
What you are suggesting is a recommender system; these can be built using a technique called collaborative filtering.
Related
It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.
Closed 9 years ago.
I have 130 objects.Each object is defined by 13 points(2-d points),these 13 points form data_unit. Thus there are 130 data_units. I want to classify these data_units into 4 classes. How can we do this.k-means is not possible in this scenario what are the alternatives.
There is a whole set of classification methods based on technique called machine learning. The ones implemented in OpenCV are described here. You can try for example Support Vector Machines. Its a nice and fairly easy in use method, with some tricks to get past data that cannot be linearly separated.
It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.
Closed 9 years ago.
I have numeric data which has missing values. I want to classify the data using Naive Bayes Classifier in Weka but the option is grayed out. Please help.
If a classifier is grayed out, it means that it doesn't support the kind of task you want to do. This could be because of the type of features or because of the type of predictions you want to make.
In your case it could be that you're trying to predict a numeric instead of a categorical attribute.
It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.
Closed 10 years ago.
I don't know if the question is well suited for stackoverflow but also it's hard to fit it to image processing.
Question is about lighting setups in computer vision systems.
Can someone provide reference to materials about lighting setups. When to use backlight, darkfield etc. How to improve features detection by lighting tricks.
I know that good lighting makes image-processing easier and want to get familiar with theory about it.
Transform your image into the Lab color space. The l channel represents the lighting.
It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.
Closed 10 years ago.
I have installed mahout and test run several examples; would it be safe to say that one of Mahout's advantages it being used for unstructured data (scalability aside)?
No... most of the algorithms take fairly specifically structured input like vectors or user-item pairs. There are some tools for vectorizing text docs which are unstructured. But no I wouldn't call it a tool primarily for unstructured data.
It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.
Closed 10 years ago.
I want to know the list of algorithms which I can do experiment to predict the probability of cancer or fever or whatever in patient based from set of inputs...Please assume that I have data in millions so I want to try the best algorithms to predict that...I am really new to data mining and machine learning....
One of the current and most popular algorithms for prediction and classification is Random Forests (RF) by Leo Breiman. Its implementation is available in weka, too.
If you are looking specifically at estimating probabilities of something, then you need to use a machine learning approach that generates probabilities. Most only generate a class label: yes/no.
The most well-known algorithm for estimating probabilities is Logistic Regression. An implementation is available in Weka.
The question being a bit vague, I can only give a vague answer : use the almighty SVM! Feed the SVM classifier with your millions of input vectors, and it should be able to give you state-of-the-art predictions afterwards.
If you're looking for an implementation of SVM, have a look at libsvm, which has wrappers in almost every decent programming language.
Most popular tool for starting experiment with large amount of machine learning libs is Weka. Here you can upload your data and try many algorithms. It's weaknes is scalability, but it is not problem for plaing with data.