As stated in the title, I have read multiple sources that Policy Gradient methods are RL are suitable for large actions spaces, but I dont understand why is this so.
I am trying to see if RL can tackle a problem of mine that has a huge combinatorial no. of possible actions. Hypothetically it is about sending n no. of items from location i to j. Any combination of (i,j,n) is a possible action, and (i,j,n) all have magnitude in the 1000s, this makes more than a billion possible actions.
Since the output layer nodes of the neural net in Policy Gradient methods represents the no. of actions. With >1000,000,000 possible actions, how can Policy Gradient be a good method to solve such problems?
For large or continuous action spaces, you need to use function approximation methods to approximate the optimal policy. This is known as policy approximation. There are a number of possible approaches that include least-squares optimization or gradient-based optimization. Nearly all of these techniques utilize random sampling to produce and compare possible actions that maximize return over the infinite-time horizon.
From Sutton and Barto's RL book 1:
Policy-based methods offer practical ways of dealing with large action spaces, even continuous spaces with an infinite number of actions. Instead of computing learned probabilities for each of the many actions, we instead learn statistics of the probability distribution. For example, the action set might be the real numbers, with actions chosen from a normal (Gaussian) distribution.
Check out:
Section 13.7 in Sutton and Barto's book for more theoretical explanation
This GitHub repo for code examples that have viable approaches for this problem
Related
I am new to NLP and feature extraction, i wish to create a machine learning model that can determine the sentiment of stock related social media posts. For feature extraction of my dataset I have opted to use Word2Vec. My question is:
Is it important to train my word2vec model on a corpus of stock related social media posts - the datasets that are available for this are not very large. Should I just use a much larger pretrained word vector ?
The only way to to tell what will work better for your goals, within your constraints of data/resources/time, is to try alternate approaches & compare the results on a repeatable quantititave evaluation.
Having training texts that are properly representative of your domain-of-interest can be quite important. You may need your representation of the word 'interest', for example, to represent that of stock/financial world, rather than the more general sense of the word.
But quantity of data is also quite important. With smaller datasets, none of your words may get great vectors, and words important to evaluating new posts may be missing or of very-poor quality. In some cases taking some pretrained set-of-vectors, with its larger vocabulary & sharper (but slightly-mismatched to domain) word-senses may be a net help.
Because these pull in different directions, there's no general answer. It will depend on your data, goals, limits, & skills. Only trying a range of alternative approaches, and comparing them, will tell you what should be done for your situation.
As this iterative, comparative experimental pattern repeats endlessly as your projects & knowledge grow – it's what the experts do! – it's also important to learn, & practice. There's no authority you can ask for any certain answer to many of these tradeoff questions.
Other observations on what you've said:
If you don't have a large dataset of posts, and well-labeled 'ground truth' for sentiment, your results may not be good. All these techniques benefit from larger training sets.
Sentiment analysis is often approached as a classification problem (assigning texts to bins of 'positive' or 'negative' sentiment, operhaps of multiple intensities) or a regression problem (assigning texts a value on numerical scale). There are many more-simple ways to create features for such processes that do not involve word2vec vectors – a somewhat more-advanced technique, which adds complexity. (In particular, word-vectors only give you features for individual words, not texts of many words, unless you add some other choices/steps.) If new to the sentiment-analysis domain, I would recommend against starting with word-vector features. Only consider adding them later, after you've achieved some initial baseline results without their extra complexity/choices. At that point, you'll also be able to tell if they're helping or not.
Terminology:
Component: PC
loading-score[i,j]: the j feature in PC[i]
Question:
I know the question regarding feature selection is asked several times here at StackOverflow (SO) and on other tech-pages, and it proposes different answers/discussion. That is why I want to open a discussion for the different solutions, and not post it as a general question since that has been done.
Different methods are proposed for feature selection using PCA: For instance using the dot product between the original features and the components (here) to get their correlation, a discussion at SO here suggests that you can only talk about important features as loading-scores in a component (and not use that importance in the input space), and another discussion at SO (which I cannot find at the moment) suggest that the importance for feature[j] would be abs(sum(loading_score[:,j]) i.e the sum of the absolute value of loading_score[i,j] for all i components.
I personally would think that a way to get the importance of a feature would be an absolute sum where each loading_score[i,j] is weighted by the explained variance of component i i.e
imp_feature[j]=sum_i (abs(loading_score[i,j])*explained_variance[i].
Well, there is no universal way to select features; it totally depends on the dataset and some insights available about the dataset. I will provide some examples which might be helpful.
Since you asked about PCA, initially it separates the whole dataset into two sets under which the variances. On the other ICA (Independent Component Analysis) is able to extract multiple features simultaneously. Look at this example,
In this example, we mix three independent signals and try to separate out them using ICA and PCA. In this case, ICA is doing it a better way than PCA. In general, if you search Blind Souce Separation (BSS) you may find more information about this. Besides, in this example, we know the independent components thus, separation is easy. In general, we do not know the number of components. Hence, you may have to guess based on some prior information about the dataset. Also, you may use LDA (Linear Discriminate Analysis) to reduce the number of features.
Once you extract PC components using any of the techniques, following way we can visualize it. If we assume, those extracted components as random variables i.e., x, y, z
More information about you may refer to this original source where I took about two figures.
Coming back to your proposition,
imp_feature[j]=sum_i (abs(loading_score[i,j])*explained_variance[i]
I would not recommend this way due to the following reasons:
abs(loading_score[i,j]) when we get absolute values you may loose positive or negative correlations of considered features. explained_variance[i] may be used to find the correlation between features, but multiplying does not make any sense.
Edit:
In PCA, each component has its explained variance. Explained variance is the ratio between individual component variance and total variance(sum of all individual components variances). Feature significance can be measured by magnitude of explained variance.
All in all, what I want to say, feature selection totally depends on the dataset and the significance of features. PCA is just one technique. Frist understand the properties of features and the dataset. Then, try to extract features. Hope this helps. If you can provide us with an exact example, we may provide more insights.
I'm performing an experiment in which I need to compare classification performance of several classification algorithms for spam filtering, viz. Naive Bayes, SVM, J48, k-NN, RandomForests, etc. I'm using the WEKA data mining tool. While going through the literature I came to know about various dimension reduction methods which can be broadly classified into two types-
Feature Reduction: Principal Component Analysis, Latent Semantic Analysis, etc.
Feature Selection: Chi-Square, InfoGain, GainRatio, etc.
I have also read this tutorial of WEKA by Jose Maria in his blog: http://jmgomezhidalgo.blogspot.com.es/2013/02/text-mining-in-weka-revisited-selecting.html
In this blog he writes, "A typical text classification problem in which dimensionality reduction can be a big mistake is spam filtering". So, now I'm confused whether dimensionality reduction is of any use in case of spam filtering or not?
Further, I have also read in the literature about Document Frequency and TF-IDF as being one of feature reduction techniques. But I'm not sure how does it work and come into play during classification.
I know how to use weka, chain filters and classifiers, etc. The problem I'm facing is since I don't have enough idea about feature selection/reduction (including TF-IDF) I am unable to decide how and what feature selection techniques and classification algorithms I should combine to make my study meaningful. I also have no idea about optimal threshold value that I should use with chi-square, info gain, etc.
In StringToWordVector class, I have an option of IDFTransform, so does it makes sence to set it to TRUE and also use a feature selection technique, say InfoGain?
Please guide me and if possible please provide links to resources where I can learn about dimension reduction in detail and can plan my experiment meaningfully!
Well, Naive Bayes seems to work best for spam filtering, and it doesn't play nicely with dimensionality reduction.
Many dimensionality reduction methods try to identify the features of the highest variance. This of course won't help a lot with spam detection, you want discriminative features.
Plus, there is not only one type of spam, but many. Which is likely why naive Bayes works better than many other methods that assume there is only one type of spam.
In our program we use a genetic algorithm since years to sole problems for n variables, each having a fixed set of m possible values. This typically works well for ~1,000 variables and 10 possibilities.
Now i have a new task where only two possibilities (on/off) exist for each variable, but i'll probably need to solve systems with 10,000 or more variables. The existing GA does work but the solution improves only very slowly.
All the EA i find are designed rather for continuous or integer/float problems. Which one is best suited for binary problems?
Well, the Genetic Algorithm in its canonical form is among the best suited metaheuristics for binary decision problems. The default configuration that I would try is such a genetic algorithm that uses 1-elitism and that is configured with roulette-wheel selection, single point crossover (100% crossover rate) and bit flip mutation (e.g. 5% mutation probability). I would suggest you try this combination with a modest population size (100-200). If this does not work well, I would suggest to increase the population size, but also change the selection scheme to a tournament selection scheme (start with binary tournament selction and increase the tournament group size if you need even more selection pressure). The reason is that with a higher population size, the fitness-proportional selection scheme might not excert the necessary amount of selection pressure to drive the search towards the optimal region.
As an alternative, we have developed an advanced version of the GA and termed it Offspring Selection Genetic Algorithm. You can also consider trying to solve this problem with a trajectory-based algorithm like Tabu Search or Simulated Annealing that just uses mutation to move from one solution to another by just making small changes.
We have a GUI-driven software (HeuristicLab) that allows you to experiment with a number of metaheuristics on several problems. Your problem is unfortunately not included, but it's GPL licensed and you can implement your own problem there (through just the GUI even, there's a howto for that).
Like DonAndre said, canonical GA was pretty much designed for binary problems.
However...
No evolutionary algorithm is in itself a magic bullet (unless it has billions of years runtime). What matters most is your representation, and how that interacts with your mutation and crossover operators: together, these define the 'intelligence' of what is essentially a heuristic search in disguise. The aim is for each operator to have a fair chance of producing offspring with similar fitness to the parents, so if you have domain-specific knowledge that allows you to do better than randomly flipping bits or splicing bitstrings, then use this.
Roulette and tournament selection and elitism are good ideas (maybe preserving more than 1, it's a black art, who can say...). You may also benefit from adaptive mutation. The old rule of thumb is that 1/5 of offspring should be better than the parents - keep track of this quantity and vary the mutation rate appropriately. If offspring are coming out worse then mutate less; if offspring are consistently better then mutate more. But the mutation rate needs an inertia component so it doesn't adapt too rapidly, and as with any metaparameter, setting this is something of a black art. Good luck!
Why not try a linear/integer program?
Closed. This question is off-topic. It is not currently accepting answers.
Want to improve this question? Update the question so it's on-topic for Stack Overflow.
Closed 10 years ago.
Improve this question
I have a few sets of questions regarding outlier detection:
Can we find outliers using k-means and is this a good approach?
Is there any clustering algorithm which does not accept any input from the user?
Can we use support vector machine or any other supervised learning algorithm for outlier detection?
What are the pros and cons of each approach?
I will limit myself to what I think is essential to give some clues about all of your questions, because this is the topic of a lot of textbooks and they might probably be better addressed in separate questions.
I wouldn't use k-means for spotting outliers in a multivariate dataset, for the simple reason that the k-means algorithm is not built for that purpose: You will always end up with a solution that minimizes the total within-cluster sum of squares (and hence maximizes the between-cluster SS because the total variance is fixed), and the outlier(s) will not necessarily define their own cluster. Consider the following example in R:
set.seed(123)
sim.xy <- function(n, mean, sd) cbind(rnorm(n, mean[1], sd[1]),
rnorm(n, mean[2],sd[2]))
# generate three clouds of points, well separated in the 2D plane
xy <- rbind(sim.xy(100, c(0,0), c(.2,.2)),
sim.xy(100, c(2.5,0), c(.4,.2)),
sim.xy(100, c(1.25,.5), c(.3,.2)))
xy[1,] <- c(0,2) # convert 1st obs. to an outlying value
km3 <- kmeans(xy, 3) # ask for three clusters
km4 <- kmeans(xy, 4) # ask for four clusters
As can be seen in the next figure, the outlying value is never recovered as such: It will always belong to one of the other clusters.
One possibility, however, would be to use a two-stage approach where one's removing extremal points (here defined as vector far away from their cluster centroids) in an iterative manner, as described in the following paper: Improving K-Means by Outlier Removal (Hautamäki, et al.).
This bears some resemblance with what is done in genetic studies to detect and remove individuals which exhibit genotyping error, or individuals that are siblings/twins (or when we want to identify population substructure), while we only want to keep unrelated individuals; in this case, we use multidimensional scaling (which is equivalent to PCA, up to a constant for the first two axes) and remove observations above or below 6 SD on any one of say the top 10 or 20 axes (see for example, Population Structure and Eigenanalysis, Patterson et al., PLoS Genetics 2006 2(12)).
A common alternative is to use ordered robust mahalanobis distances that can be plotted (in a QQ plot) against the expected quantiles of a Chi-squared distribution, as discussed in the following paper:
R.G. Garrett (1989). The chi-square plot: a tools for multivariate outlier recognition. Journal of Geochemical Exploration 32(1/3): 319-341.
(It is available in the mvoutlier R package.)
It depends on what you call user input. I interpret your question as whether some algorithm can process automatically a distance matrix or raw data and stop on an optimal number of clusters. If this is the case, and for any distance-based partitioning algorithm, then you can use any of the available validity indices for cluster analysis; a good overview is given in
Handl, J., Knowles, J., and Kell, D.B.
(2005). Computational cluster validation in post-genomic data analysis.
Bioinformatics 21(15): 3201-3212.
that I discussed on Cross Validated. You can for instance run several instances of the algorithm on different random samples (using bootstrap) of the data, for a range of cluster numbers (say, k=1 to 20) and select k according to the optimized criteria taht was considered (average silhouette width, cophenetic correlation, etc.); it can be fully automated, no need for user input.
There exist other forms of clustering, based on density (clusters are seen as regions where objects are unusually common) or distribution (clusters are sets of objects that follow a given probability distribution). Model-based clustering, as it is implemented in Mclust, for example, allows to identify clusters in a multivariate dataset by spanning a range of shape for the variance-covariance matrix for a varying number of clusters and to choose the best model according to the BIC criterion.
This is a hot topic in classification, and some studies focused on SVM to detect outliers especially when they are misclassified. A simple Google query will return a lot of hits, e.g. Support Vector Machine for Outlier Detection in Breast Cancer Survivability Prediction by Thongkam et al. (Lecture Notes in Computer Science 2008 4977/2008 99-109; this article includes comparison to ensemble methods). The very basic idea is to use a one-class SVM to capture the main structure of the data by fitting a multivariate (e.g., gaussian) distribution to it; objects that on or just outside the boundary might be regarded as potential outliers. (In a certain sense, density-based clustering would perform equally well as defining what an outlier really is is more straightforward given an expected distribution.)
Other approaches for unsupervised, semi-supervised, or supervised learning are readily found on Google, e.g.
Hodge, V.J. and Austin, J. A Survey of Outlier Detection Methodologies.
Vinueza, A. and Grudic, G.Z. Unsupervised Outlier Detection and Semi-Supervised Learning.
Escalante, H.J. A Comparison of Outlier Detection Algorithms for Machine Learning.
A related topic is anomaly detection, about which you will find a lot of papers.
That really deserves a new (and probably more focused) question :-)
1) Can we find outliers using k-means, is it a good approach?
Cluster-based approaches are optimal to find clusters, and can be used to detect outliers as
by-products. In the clustering processes, outliers can affect the locations of the cluster centers, even aggregating as a micro-cluster. These characteristics make the cluster-based approaches infeasible to complicated databases.
2) Is there any clustering algorithm which does not accept any input from the user?
Maybe you can achieve some valuable knowledge on this topic:
Dirichlet Process Clustering
Dirichlet-based clustering algorithm can adaptively determine the number of clusters according to the distribution of observation data.
3) Can we use support vector machine or any other supervised learning algorithm for outlier detection?
Any Supervised learning algorithm needs enough labeled training data to construct classifiers. However, a balanced training dataset is not always available for real world problem, such as intrusion detection, medical diagnostics. According to the definition of Hawkins Outlier("Identification of Outliers". Chapman and Hall, London, 1980), the number of normal data is much larger than that of outliers. Most supervised learning algorithms can't achieve an efficient classifier on the above unbalanced dataset.
4) What is the pros and cons of each approach?
Over the past several decades, the research on outlier detection varies from the global computation to the local analysis, and the descriptions of outliers vary from the binary interpretations to probabilistic representations. According to hypotheses of outlier detection models, outlier detection algorithms can be divided into four kinds: Statistic-based algorithms, Cluster-based algorithms, Nearest Neighborhood based algorithms, and Classifier-based algorithms. There are several valuable surveys on outlier detection:
Hodge, V. and Austin, J. "A survey of outlier detection methodologies", Journal of Artificial Intelligence Review, 2004.
Chandola, V. and Banerjee, A. and Kumar, V. "Outlier detection: A survey", ACM Computing Surveys, 2007.
k-means is rather sensitive to noise in the data set. It works best when you remove the outliers beforehand.
No. Any cluster analysis algorithm that claims to be parameter-free usually is heavily restricted, and often has hidden parameters - a common parameter is the distance function, for example. Any flexible cluster analysis algorithm will at least accept a custom distance function.
one-class classifiers are a popular machine-learning approach to outlier detection. However, supervised approaches aren't always appropriate for detecting _previously_unseen_ objects. Plus, they can overfit when the data already contains outliers.
Every approach has its pros and cons, that is why they exist. In a real setting, you will have to try most of them to see what works for your data and setting. It's why outlier detection is called knowledge discovery - you have to explore if you want to discover something new ...
You may want to have a look at the ELKI data mining framework. It is supposedly the largest collection of outlier detection data mining algorithms. It's open source software, implemented in Java, and includes some 20+ outlier detection algorithms. See the list of available algorithms.
Note that most of these algorithms are not based on clustering. Many clustering algorithms (in particular k-means) will try to cluster instances "no matter what". Only few clustering algorithms (e.g. DBSCAN) actually consider the case that maybe not all instance belong into clusters! So for some algorithms, outliers will actually prevent a good clustering!