Is there any rule of thumb to initialize the num_leaves parameter in lightgbm. For example for 1000 featured dataset, we know that with tree-depth of 10, it can cover the entire dataset, so we can choose this accordingly, and search space for tuning also get limited.
But in lightgbm, how we can roughly guess this parameters, otherwise its search space will be pretty large while using grid-search method.
Any intuition on selecting this parameters will be helpful.
The best recommendation, that I bumped into is this awesome summary by Laurae on lightgbm github. As always, this very much depends on your data.
My personal rule of thumb based on limited kaggle experience is to start by trying values in the range [10,100]. But if you have a solid heuristic to choose tree depth you can always use it and set num_leaves to 2^tree_depth - 1
Related
I need some point of view to know if what I am doing is good or wrong or if there is better way to do it.
I have 10 000 elements. For each of them I have like 500 features.
I am looking to measure the separability between 2 sets of those elements. (I already know those 2 groups I don't try to find them)
For now I am using svm. I train the svm on 2000 of those elements, then I look at how good the score is when I test on the 8000 other elements.
Now I would like to now which features maximize this separation.
My first approach was to test each combination of feature with the svm and follow the score given by the svm. If the score is good those features are relevant to separate those 2 sets of data.
But this takes too much time. 500! possibility.
The second approach was to remove one feature and see how much the score is impacted. If the score changes a lot that feature is relevant. This is faster, but I am not sure if it is right. When there is 500 feature removing just one feature don't change a lot the final score.
Is this a correct way to do it?
Have you tried any other method ? Maybe you can try decision tree or random forest, it would give out your best features based on entropy gain. Can i assume all the features are independent of each other. if not please remove those as well.
Also for Support vectors , you can try to check out this paper:
http://axon.cs.byu.edu/Dan/778/papers/Feature%20Selection/guyon2.pdf
But it's based more on linear SVM.
You can do statistical analysis on the features to get indications of which terms best separate the data. I like Information Gain, but there are others.
I found this paper (Fabrizio Sebastiani, Machine Learning in Automated Text Categorization, ACM Computing Surveys, Vol. 34, No.1, pp.1-47, 2002) to be a good theoretical treatment of text classification, including feature reduction by a variety of methods from the simple (Term Frequency) to the complex (Information-Theoretic).
These functions try to capture the intuition that the best terms for ci are the
ones distributed most differently in the sets of positive and negative examples of
ci. However, interpretations of this principle vary across different functions. For instance, in the experimental sciences χ2 is used to measure how the results of an observation differ (i.e., are independent) from the results expected according to an initial hypothesis (lower values indicate lower dependence). In DR we measure how independent tk and ci are. The terms tk with the lowest value for χ2(tk, ci) are thus the most independent from ci; since we are interested in the terms which are not, we select the terms for which χ2(tk, ci) is highest.
These techniques help you choose terms that are most useful in separating the training documents into the given classes; the terms with the highest predictive value for your problem. The features with the highest Information Gain are likely to best separate your data.
I've been successful using Information Gain for feature reduction and found this paper (Entropy based feature selection for text categorization Largeron, Christine and Moulin, Christophe and Géry, Mathias - SAC - Pages 924-928 2011) to be a very good practical guide.
Here the authors present a simple formulation of entropy-based feature selection that's useful for implementation in code:
Given a term tj and a category ck, ECCD(tj , ck) can be
computed from a contingency table. Let A be the number
of documents in the category containing tj ; B, the number
of documents in the other categories containing tj ; C, the
number of documents of ck which do not contain tj and D,
the number of documents in the other categories which do
not contain tj (with N = A + B + C + D):
Using this contingency table, Information Gain can be estimated by:
This approach is easy to implement and provides very good Information-Theoretic feature reduction.
You needn't use a single technique either; you can combine them. Term-Frequency is simple, but can also be effective. I've combined the Information Gain approach with Term Frequency to do feature selection successfully. You should experiment with your data to see which technique or techniques work most effectively.
If you want a single feature to discriminate your data, use a decision tree, and look at the root node.
SVM by design looks at combinations of all features.
Have you thought about Linear Discriminant Analysis (LDA)?
LDA aims at discovering a linear combination of features that maximizes the separability. The algorithm works by projecting your data in a space where the variance within classes is minimum and the one between classes is maximum.
You can use it reduce the number of dimensions required to classify, and also use it as a linear classifier.
However with this technique you would lose the original features with their meaning, and you may want to avoid that.
If you want more details I found this article to be a good introduction.
I am going to use ANN for my work in which I have a large dataset, let say input[600x40] and output[600x6]. As one can see, the number of inputs (40) is too high for ANN and it may trap in local minimum and/or increases the CPU time dramatically. Is there any way to select the most informative input?
As my first try, I used the following code in Matlab to find the cross-correlation between each two inputs:
[rho, ~] = corr(inputs, 'rows','pairwise')
However, I think this simple correlation cannot identify some hidden complex relation between the inputs.
Any ideas?
First of all 40 inputs is a very small space and it should not be reduced. Large number of inputs is 100,000, not 40. Also, 600x40 is not a big dataset, nor the one "increasing the CPU time dramaticaly", if it learns slowly than check your code because it appears to be the problem, not your data.
Furthermore, feature selection is not a good way to go, you should use it only when gathering features is actually expensive. In any other scenario you are looking for dimensionality reduction, such as PCA, LDA etc. although as said before - your data should not be reduced, rather - you should consider getting more of it (new samples/new features).
Disclaimer: I'm with lejlot on this - you should get more data and
more features instead of trying to remove features. Still, that doesn't answer your question, so here we go.
Try most basic greedy approach - try removing each feature and retrain your ANN (several times, of course) and see if your results got better or worse. Choose this situation where results got better and improvement was the best. Repeat until you'll get no improvement by removing features. This will take a lot of time, so you may want to try doing it on some subset of your data (for example on 3 folds of dataset splitted into 10 folds).
It's ugly, but sometimes it works.
I repeat what I've said in disclaimer - this is not the way to go.
I have a large image (5400x3600) that has multiple CCTVs that I need to detect.
The detection takes lot of time (4-7 minutes) with rotation. But it still fails to resolve certain CCTVs.
What is the best method to match a template like this?
I am using skImage - openCV is not an option for me, but I am open to suggestions on that too.
For example: in the images below, the template is correct matched with the second image - but the first image is not matched - I guess due to the noise created by the text "BLDG..."
Template:
Source image:
Match result:
The fastest method is probably a cascade of boosted classifiers trained with several variations of your logo and possibly a few rotations and some negative examples too (non-logos). You have to roughly scale your overall image so the test and training examples are approximately matched by scale. Unlike SIFT or SURF that spend a lot of time in searching for interest points and creating descriptors for both learning and searching, binary classifiers shift most of the burden to a training stage while your testing or search will be much faster.
In short, the cascade would run in such a way that a very first test would discard a large portion of the image. If the first test passes the others will follow and refine. They will be super fast consisting of just a few intensity comparison in average around each point. Only a few locations will pass the whole cascade and can be verified with additional tests such as your rotation-correlation routine.
Thus, the classifiers are effective not only because they quickly detect your object but because they can also quickly discard non-object areas. To read more about boosted classifiers see a following openCV section.
This problem in general is addressed by Logo Detection. See this for similar discussion.
There are many robust methods for template matching. See this or google for a very detailed discussion.
But from your example i can guess that following approach would work.
Create a feature for your search image. It essentially has a rectangle enclosing "CCTV" word. So the width, height, angle, and individual character features for matching the textual information could be a suitable choice. (Or you may also use the image having "CCTV". In that case the method will not be scale invariant.)
Now when searching first detect rectangles. Then use the angle to prune your search space and also use image transformation to align the rectangles in parallel to axis. (This should take care of the need for the rotation). Then according to the feature choosen in step 1, match the text content. If you use individual character features, then probably your template matching step is essentially a classification step. Otherwise if you use image for matching, you may use cv::matchTemplate.
Hope it helps.
Symbol spotting is more complicated than logo spotting because interest points work hardly on document images such as architectural plans. Many conferences deals with pattern recognition, each year there are many new algorithms for symbol spotting so giving you the best method is not possible. You could check IAPR conferences : ICPR, ICDAR, DAS, GREC (Workshop on Graphics Recognition), etc. This researchers focus on this topic : M Rusiñol, J Lladós, S Tabbone, J-Y Ramel, M Liwicki, etc. They work on several techniques for improving symbol spotting such as : vectorial signatures, graph based signature and so on (check google scholar for more papers).
An easy way to start a new approach is to work with simples shapes such as lines, rectangles, triangles instead of matching everything at one time.
Your example can be recognized by shape matching (contour matching), much faster than 4 minutes.
For good match , you require nice preprocess and denoise.
examples can be found http://www.halcon.com/applications/application.pl?name=shapematch
I'm working on an algorithm that makes a guess at K for kmeans clustering. I guess I'm looking for a data set that I could use as a comparison, or maybe a few data sets where the number of clusters is "known" so I could see how my algorithm is doing at guessing K.
I would first check the UCI repository for data sets:
http://archive.ics.uci.edu/ml/datasets.html?format=&task=clu&att=&area=&numAtt=&numIns=&type=&sort=nameUp&view=table
I believe there are some in there with the labels.
There are text clustering data sets that are frequently used in papers as baselines, such as 20newsgroups:
http://qwone.com/~jason/20Newsgroups/
Another great method (one that my thesis chair always advocated) is to construct your own small example data set. The best way to go about this is to start small, try something with only two or three variables that you can represent graphically, and then label the clusters yourself.
The added benefit of a small, homebrew data set is that you know the answers and it is great for debugging.
Since you are focused on k-means, have you considered using the various measures (Silhouette, Davies–Bouldin etc.) to find the optimal k?
In reality, the "optimal" k may not be a good choice. Most often, one does want to choose a much larger k, then analyze the resulting clusters / prototypes in more detail to build clusters out of multiple k-means partitions.
The iris flower dataset is a good one to start with, that clustering works nicely on.
Download here
Consider the text classification problem of spam or not spam with the Naive Bayes algorithm.
The question is the following:
how do you make predictions about a document W = if in that set of words you see a new word wordX that was not seen at all by your model (so you do not even have a laplace smoothing probabilty estimated for it)?
Is the usual thing to do is just ignore that wordX eventhough it was seen in the current text because it has no probability associated with? I.e. I know sometimes the laplace smoothing is used to try to solve this problem, but what if that word is definitively new?
Some of the solutions that I've thought of:
1) Just ignore that words in estimating a classification (most simple, but sometimes wrong...?, however, if the training set is large enough, this is probably the best thing to do, as I think its reasonable to assume your features and stuff were selected well enough if you have say 1M or 20M data).
2) Add that word to your model and change your model completely, because the vocabulary changed so probabilities have to change everywhere (this does have a problem though since it could mean that you have to update the model frequently, specially if your analysis 1M documents, say)
I've done some research on this, read some of the Dan Jurafsky NLP and NB slides and watched some videos on coursera and looked through some research papers but I was not able to find something I found useful. It feels to me this problem is not new at all and there should be something (a heuristic..?) out there. If there isn't, it would be awesome to know that too!
Hope this is a useful post for the community and Thanks in advance.
PS: to make the issue a little more explicit with one of the solutions I've seen is, say that we see an unknown new word wordX in a spam, then for that word we can do 1/ count(spams) + |Vocabulary + 1|, the issue I have with doing something like that is that, then, does that mean we change the size of the vocabulary and now, every new document we classify, has a new feature and vocabulary word? This video seems to attempt to solve that issue but I'm not sure if either, thats a good thing to do or 2, maybe I have misunderstood it:
https://class.coursera.org/nlp/lecture/26
From a practical perspective (keeping in mind this is not all you're asking), I'd suggest the following framework:
Train a model using an initial train set, and start using it for classificaion
Whenever a new word (with respect to your current model) appears, use some smoothing method to account for it. e.g. Laplace smoothing, as suggested in the question, might be a good start.
Periodically retrain your model using new data (usually in addition to the original train set), to account for changes in the problem domain, e.g. new terms. This can be done on preset intervals, e.g once a month; after some number of unknown words was encountered, or in an online manner, i.e. after each input document.
This retrain step can be done manually, e.g. collect all documents containing unknown terms, manually label them, and retrain; or using semi-supervised learning methods, e.g. automatically add the highest scored spam/ non spam documents to the respective models.
This will ensure your model stays updated and accounts for new terms - by adding them to the model from time to time, and by accounting for them even before that (simply ignoring them is usually not a good idea).