Questions tagged [unsupervised-learning]

Finding hidden (statistical) structure in unlabelled data, including clustering and feature extraction for dimensionality reduction.

3
votes
1answer
249 views

Assumptions on class priors in expectation-maximization?

I want to use the EM algorithm to do clustering under a missing labels regime. The assumption I am making about the missing data is that it's distributed according to Bernoulli distribution. So for ...
0
votes
3answers
243 views

Detect and remove abnormal data from 1D data

I have a 1D data which represents order values. Sometimes a client creates a test order which they don't remove from the system. What would be the best way of detecting and removing values that seem ...
0
votes
0answers
16 views

Which machine learning algorithm may help to explore the importance of contributions of variables to their sum over time?

I have data in the form of a $n \times t$ matrix $X$ where $n$ is a number of variables and $t$ a (large) number of time points. At any given time point the elements of the matrix can be expressed as ...
7
votes
3answers
9k views

Do we need to set training set and testing set for clustering?

When we do classification and regression, we usually set testing and training sets to help us build and improve models. However, when we do clustering do we also need to set testing and training sets?...
2
votes
1answer
204 views

Empty nodes when creating SOM

I am trying to create a SOM map based on records with different discrete classifications (tags) like the example below ...
0
votes
0answers
76 views

What unsupervised methods are used for sampling a dataset?

I have a dataset of 21 classes. Each class has different number of samples. I want to sample an specific number of samples from each class. For example 50. I thought of using K-means to cluster 50 ...
-1
votes
2answers
2k views

Support Vector Machine - Classification or Clustering

I don't really understand if SVM are classification methods (like Logistic regression) or clustering methods. Since it's used for supervised learning, it should be part of classification methods ...
0
votes
1answer
557 views

Comparing supervised text classification algorithms with unlabeled documents from web

Working with the unlabeled documents from web for supervised text classification, even though the problem settings dictates using semi-supervised learning, I aim to compare several different ...
1
vote
2answers
1k views

Constrained Clustering Implementation R or Python? [closed]

Can anyone point me to an implementation in R or Python of a constrained clustering algorithm? In case this is overly broad, I am hoping to exploit known must-link/cannot-link pairs to improve the ...
1
vote
0answers
56 views

Supervised learning for audio files with uncertain labels

There exists a collection of audio files, each 30 seconds long. Such an audio file contains the recording of the manufacturing of a certain small piece in a machine. After manufacturing the piece, it ...
1
vote
0answers
36 views

How to check the quality of clustering results if there is no labels? [duplicate]

How can we check the quality of clustering of no-labeled data? I learnt from class that there are some ways to achieve this . One is to "measure intra-cluster cohesion (how near the data points ...
2
votes
2answers
817 views

Unsupervised Topic Models that don't require the number of topics to be set upfront

I hope this question isn't too general, but I'm looking for an unsupervised topic modelling algorithm that doesn't require the number of topics (k) to be defined prior to running the analysis of the ...
0
votes
1answer
151 views

Multiple correspondence analysis for clustering (unsupervised learning)

I have limited stat/coding knowledge yet I try to do user clustering using unsupervised method using R. I have about 2795 observations gained from survey (mixture of categorical and scale questions). ...
1
vote
0answers
453 views

Finding brief repeated patterns in a time series

Suppose I have a series of events, each of which is an element of {-1, 0 1} with a known distribution. Events happen in approximately continuous time. My dependent variable is the sum of the series. ...
1
vote
0answers
52 views

Does it make sense to use auto-encoders to reconstruct GIST features?

I am trying to extract good low dimensional representation of CIFAR-10 images in an unsupervised way. It is a project requirement that I use 512-d GIST features, reduce the dimensions to 32 using PCA ...
1
vote
1answer
822 views

Classification using lookup table

I have a matrix of samples to classify, samples are matrix columns and features (noisy or estimated features) are matrix rows. On the other hand, I have a lookup table for correspondence between ...
0
votes
1answer
1k views

Do RNNs (recurrent neural networks) support feature learning or not?

Please explain how RNNs support feature learning problem of deep learning or why it does not support it. How about unsupervised learning; Does RNN support that?
29
votes
5answers
5k views

Distinguishing between two groups in statistics and machine learning: hypothesis test vs. classification vs. clustering

Assume I have two data groups, labeled A and B (each containing e.g. 200 samples and 1 feature), and I want to know if they are different. I could: a) perform a statistical test (e.g. t-test) to see ...
0
votes
1answer
1k views

How to perform Validation on Unsupervised learning?

Since I consider Unsupervised learning, I don't have any ground truth to compare with, during the validation phase. So, is there any standard method to deal with it? Additional informations: in my ...
1
vote
0answers
130 views

Performance measure for iterative semi-supervised learning

Consider the problem of semi-supervised learning where, in each round, the labels of all data points are guessed and then the label of a random data point is revealed. As the labels of more and more ...
0
votes
0answers
40 views

Are there clustering algorithms which require a human as an oracle?

Are there data clustering algorithms which by definition require human judgement as part of the algorithm? What I mean is that once in a while the algorithm presents the human operator (or any other ...
0
votes
1answer
41 views

Find optimal configuration ranges for manufacturing process

Let's assume I have some industrial manufacturing process with one costs and a produced output as well as a number of variables to configure the process. Based on my historic data I would like to ...
3
votes
2answers
6k views

How to assign new data to an existing clustering

I have the following case. Say I have a set of 100 celebrities and I form 4 clusters using k-means. Lets assume that these 4 clusters are music, sports, politics, movies. Now say if I want to ...
3
votes
1answer
2k views

Sequence to sequence autoencoder - decoder input?

What's the input to the decoder part of a sequence to sequence autoencoder? I've seen certain examples of such an autoencoder (using LSTM's more often than not) but am still unclear. For example, ...
1
vote
0answers
224 views

Many 'feature selection' in the same dataset

If I have, for example, a dataset with 30 variables, 25 numerical and 5 categorical. I need to explain (predict) 3 categorical variables and 2 numerical variables from that dataset (let's call it "...
1
vote
1answer
218 views

Why does Latent Dirichlet Allocation seems to work with greedy selection but not with Gibbs sampling?

I tried to implement my own LDA program in python, while following this tutorial. When I use gibbs sampling, the program assigns all words to a particular topic on convergence. When I greedily ...
0
votes
1answer
210 views

Classification or Clustering Approach for Time Series Data of Flow

I have a dataset which contains time-series data of water flow over time. I have a flow meter connected to a kitchen faucet, and I am trying to cluster or classify specific water usage events. The ...
1
vote
3answers
381 views

What is the best way to present clustering result? [closed]

I know that for supervised classification one can use a confusion matrix to present the results. Is there an equivalent for this for clustering? And what's the best way to present clustering-...
4
votes
3answers
1k views

Tutorial for feature extraction on unsupervised learning

I would like to extract features from (without loss of generality) numerical data using unsupervised learning methods among these: transformations: PCA/ICA/NMF embeddings: T-distributed stochastic ...
4
votes
1answer
59 views

In variational inference on von Mises clusters, how to find a bound for the log-Bessel function?

This paper on von Mises clustering uses an upper bound on the modified log-Bessel function that I struggle to replicate. Taking results from this paper, the authors state: $$u\frac{I'_\nu(u)}{I_\nu(u)...
0
votes
0answers
107 views

Continuous data discretization rules

I know it all depends on the data, but I am looking for a general, most common rules for continuous data discretization. For example It could be a list like this: Use Supervised MDL discretization (...
1
vote
2answers
463 views

Lower bound for Adjusted Rand Index?

From the original paper, it's not clear whether the adjusted rand index has a lower bound. Does it? If so, what partition yields the bound? If now, how can I construct partitions with arbitrary low ...
3
votes
1answer
2k views

Variational Auto-encoders vs Restricted Boltzmann Machines

What are the differences of modeling ability between Variational Auto-encoders (VAEs) and Restricted Boltzmann Machines (RBMs)? What I am interested in is to know about the unsupervised learning ...
1
vote
0answers
42 views

Bayesian profile-based customer discovery using demographic data

Consider the problem of estimating the probability of a person being interested in becoming a customer of a service or a buyer of a product, and only data about the current customers is available, ...
5
votes
4answers
2k views

Is overfitting a problem in unsupervised learning?

Consider the density estimation problem for some training set $(x_1 ... x_N)$. A gaussian mixture model consisting of $N$ normal distributions centered on each $x_i$ with very small variances will "...
-1
votes
3answers
119 views

What's a simple data set to to try clustering algorithms on? [closed]

I'm very new to Machine Learning and programming. I'm trying to run clustering algorithms on data sets but all examples I find contains lot of data and it confuses me. Can you give me a simple sample ...
1
vote
0answers
71 views

Hamming distance calculation in the paper “unsupervised learning of visual representation by solving Jigsaw puzzle”

I am reading the paper https://arxiv.org/abs/1603.09246. Here they speak about calculating the hamming distance for the generated permutations. In algorithm 1, under the iteration part the ...
6
votes
2answers
2k views

Constructing features from k-means

I would like to construct features from the result of apply k-means clustering to construct features of my data that I can later use for a classifier. Assume I have fixed the $k$ (e.g. 5) and ...
13
votes
3answers
3k views

Choosing the hyperparameters using T-SNE for classification

In as specific problem that I work with (a competition) I have the follwoing setting: 21 features (numerical on [0,1]) and a binary output. I have approx 100 K rows. The setting seems to be very noisy....
1
vote
0answers
145 views

Unsupervised transformation with Random Trees

I just found a an algorithm (in scikit-learn) that performs Random Trees Embedding. I see that its quite powerful but I dont really understand how it works. Is there any intuitive explanation + some ...
5
votes
1answer
103 views

Combine reinforces and unsupervised learning?

I have an existing set of data and plan to generate more data that follows the same pattern. To do this, I plan to use unsupervised learning. How can I provide feedback on the generated data and ...
3
votes
1answer
2k views

Labeling documents with short text labels after topic modeling?

If I generate a topic model (LDA, PLSA) for a group of documents, is there then a way that I could label each document with a one-to-two word label that describes the document content? For example, ...
1
vote
0answers
173 views

Which algorithm is robust to noisy data? (Decision Tree, K-Mean clustering, HMM)

I assume HMM will be the most robust to noisy data since it derives a generative model compared to Decision Tree and K-Mean? Between decision tree and K-mean, which methods is robust to noisy data? I ...
4
votes
3answers
7k views

Self organizing maps vs k-means, when the SOM has a lot of nodes

On Wikipedia it says: It has been shown that while self-organizing maps with a small number of nodes behave in a way that is similar to k-means, larger self-organizing maps rearrange data in a way ...
1
vote
0answers
100 views

how to measure the discriminative power of autoencoder

The autoencoder is used for encoding features to vectors which are supposed to be more discriminative for classification or clustering. But how to measure the discriminative power of the code. Is ...
3
votes
0answers
199 views

Neural Networks: utilizing weakly-labeled data to improve fully-supervised network?

The problem: I have built a fully-supervised CNN that localizes an object in different scenes. As you can imagine, it quite time-consuming to label data: I have to manually localize the object in an ...
1
vote
0answers
13 views

Intuition-building examples to help choose the right linkage method in hierarchical clustering [duplicate]

I'm currently reading An Introduction to Statistical Learning with Applications in R by Hastie & Tibshirani. In their discussion on the hierarchical clustering algorithm, they note that the notion ...
0
votes
1answer
331 views

Machine Learning - Aggregate granular cluster predictions on denormalized data

I have a question that occurred when thinking about the following use case: A bank wants to group their customers into segments using the database tables 'customer', 'account' and 'transaction'. The ...
0
votes
0answers
63 views

Given a pdf which is a mixture of Gaussians, how do I infer the position (mean), variance, and number of Gaussians?

I have the following data, which when plotted as a histogram, are a mixture of Gaussians: I would like to write an algorithm that would infer: (1) the number of "peaks" or normal distributions in ...
0
votes
1answer
220 views

How do you infer the number of clusters in data using unsupervised learning? [duplicate]

Let's say you have a density plot of data in 2D (or even 1D). Surely there are algorithms which infer the number of clusters which exist in the data without users having to explicitly set this number (...