Methods and principles of building "computer systems that automatically improve with experience."

learn more… | top users | synonyms (1)

0
votes
1answer
16 views

How to tune parameters through cross-validation without grid search?

There are actually lots of questions about parameter tuning through cross-validation. I have read some of them, e.g. this one. I, however, still can't understand the details of the process. Here are ...
2
votes
0answers
16 views

Variational Inference: good inference but ELBO decreases instead of increasing

I am playing with Variational Inference for clustering within a mixture of Gaussians. My first implementation seems to work fine (this is for the geyser dataset): ...
0
votes
0answers
8 views

How to calculate the probability of a sample to get a specific pattern, given all samples and features?

I have a data aggregation tool, which collects information on many samples (clusters of molecules in this case). So it can create a sparse matrix of NxM binary values (N answers of 0 or 1 for M ...
0
votes
0answers
19 views

Error function of two layer neural network

Let's say we have a two layer neural network with $\theta^1$ and $\theta^2$ as the layer units. Now ${\delta J \over \delta \theta^2}$ is convex and has a global minima for each unit. But what is the ...
1
vote
2answers
40 views

Is validation set always necessary?

Lets say I did the following steps: Used some separate development set to select some features. Decided a priori to use only one learning algorithm (SVM) with only default parameter values. Trained ...
1
vote
0answers
15 views

Different Methods for clustering skills in text

Consider a talent pool in which each member has some set of skills. Some of these talent are submitted to orders as potential candidates of which one is selected. It is reasonable to assume that the ...
1
vote
2answers
51 views

How to use cross-validation with regularization?

I think I understand each of these concepts (cross-validation, regularization) independently, but I'm not quite clear on how they can be put together in practice. Loosely speaking, in ...
2
votes
1answer
42 views

How to simulate more data for machine learning?

I am attempting to analyze a small dataset using machine learning (SVM, binary problem). There are $103$ samples and $215$ variables (all variables are real numbers). Some of the variables (around ...
0
votes
0answers
20 views

Data augmentation techniques for general datasets?

In many machine learning applications, the so called data augmentation methods have allowed building better models. For example, assume a training set of $100$ images of cats and dogs. By rotating, ...
0
votes
0answers
24 views

How to model the relation between two variables based on other common relations?

I am trying to model an inference problem, but it doesn't seem to readily fit with the algorithms we usually hear about. I am hoping that I have missed something, and someone hear can point out that ...
0
votes
0answers
18 views

Does this simple anomaly detection algorithm work for > 100 features?

In Andrew Ng's Coursera class on Machine Learning, we learned to use a Gaussian distribution $p(x) = \prod_{j=1}^n p(x_j,\mu_j,\sigma_j^2)$ to detect anomalous examples when $p(x) < \epsilon$ ...
0
votes
1answer
33 views

Opportunities in machine learning and computational intelligence [on hold]

I'm not sure this is the right site to post my question. If not, please direct me to the right one. I'm interested in machine learning and computational intelligence. I've spent the last year of my ...
0
votes
0answers
29 views

Feature extraction based on correlations

I have a small problem regarding feature extraction with correlation. I have divided my question in four parts hoping that somebody can help me. I have a dataset consisting of fMRI images. Each image ...
0
votes
2answers
30 views

No regularisation term for bias unit in neural network

According to this tutorial on deep learning, weight decay (regularization) is not usually applied to the bias terms b why? What is significance (intuition) behind it?
3
votes
2answers
40 views

Recurrent vs Recursive Neural Networks: Which applies better for NLP?

So, we have Recurrent Neural Networks and Recursive Neural Networks. Both are usually denoted by the same acronym: RNN. According to Wikipedia, Recurrent NN are in fact Recursive NN, but I don't ...
0
votes
1answer
32 views

Neural Networks and Numeric Prediction

I'm new to machine learning and am trying to write a simple neural network that uses back-propagation. Now, so far I've successfully implemented my neural network to learn a boolean function. So for ...
2
votes
0answers
32 views

Estimating probability distribution function of data stream

Although a similar question exists, I couldn't find my answer. I'm not a statistician hence please neglect if some terminologies aren't correct and let me know if I am interpreting something wrong. ...
2
votes
0answers
32 views

How to find a cost function with only a statistical measure of success?

Using the U.S.A. as a loose analogy, we have search algorithms that find the names and number of States adjacent to a given State (containing a selected city). The goal is to minimize the number of ...
4
votes
2answers
70 views

Sequential pattern mining on single sequence

Can someone give me a hint about a good approach to find a frequent patterns in a single sequence. For example there is the single sequence ...
0
votes
1answer
20 views

Intuition behind sparsity in over-complete sparse auto-encoders

I am trying to get a grasp of the intuition behind the sparse representation used in over-complete auto-encoders. One piece of text that offers a somewhat intuitive explanation is from ...
0
votes
0answers
8 views

Interpreting the lift curve

Suppose we have two classes: A and B. Suppose we use a logistic regression to assign each unit to A or B. The curve lift is calculated through this formula: $\frac{n_{22}/n_{.2}}{n_{2.}/n_{..}}$ ...
1
vote
1answer
12 views

Is there a version of Latent Class Analysis with unspecified # of clusters

I understand that you can use the elbow method to plot LCA solutions vs log likelihood to figure out, at which k, it is no longer worth it to add more clusters. And I will resort to this if need be. ...
0
votes
0answers
55 views
+50

Recurrent Neural Network (LSTM) Questions

I am coding an RNN-LSTM module, and I have the following questions: 1) In non-linear input like one-hot-encoded words, you use an extra sign (like <eol>) at ...
1
vote
0answers
17 views

Measuring effectiveness of marketing through attribution analysis [on hold]

My data(dataframe in R) looks like this:The data is ordered by CustomerName and then TimeofEvent. ...
2
votes
1answer
39 views

Predicting the job switching period of an employee

I have a machine learning problem to solve. Given the data about employees, is there a way by which we could possibly predict that when an employee is going to switch his current job? We can make use ...
0
votes
1answer
38 views

Is building training data set from unlabeled data considered as a scientific contribution? [closed]

Is building a training data set, from unlabeled data, for a machine learning classifier considered as a scientific contribution?
0
votes
1answer
30 views

What are the differences between autoencoder and t-SNE?

As far as I know, both autoencoder and t-SNE are used for nonlinear dimension reduction. What are the differences between them and why should I use one versus another? thanks!
5
votes
2answers
104 views

What is bucketization?

I've been going around to find a clear explanation of "bucketization" in machine learning with no luck. What I understand so far is that bucketization is similar to quantization in digital signal ...
0
votes
1answer
32 views

GridSearchCV Regression vs Linear Regression vs Stats.model OLS

I am trying to build multiple linear regression model with 3 different method and I am getting different results for each one. I think that I have to get the same results but ...
2
votes
1answer
43 views
+150

Standardization for regularized, sparse hashed logistic regression

As the question states, I'm fitting large, sparse logistic regressions (with hashed interactions, a la vowpal wabbit) for a machine learning system. The features are on different scales, and I'm a ...
3
votes
0answers
30 views

Markov decision process in R for a song suggestion software?

We have a music player that has different playlists and automatically suggests songs from the current playlist I'm in. What I want the program to learn is, that if I skip the song, it should decrease ...
6
votes
1answer
83 views

Bag-of-Words for Text Classification: Why not just use word frequencies instead of TFIDF?

A common approach to text classification is to train a classifier off of a 'bag-of-words'. The user takes the text to be classified and counts the frequencies of the words in each object, followed by ...
3
votes
0answers
27 views

Variance of binomial vs. multinomial distribution in cross-validation

Suppose we have a dataset with $N=100$ observations. We do $K$-fold cross-validation with $K=10$ and $K=100$. In the first case, the classification decisions are sampled (can I say it like this?) ...
2
votes
1answer
20 views

ROC and constant factor on probabilities

I play around with a few data to learn and I am wondering about something; I can evaluate my results with ROC which is processed from FP and FN. I had predicted a few probabilities for my events to ...
0
votes
0answers
40 views

Finding algorithm to detect anomaly in non gaussian data

I have a data (time series like CPU, traffic and so on) that doesnt have a normal distribution usually (especially when I'm looking at 1 hour data). Are there any algorithm to find anomalies? I ...
0
votes
1answer
46 views

Why the majority of serious machine learners are from the computer vision society? [closed]

I have noticed that most individuals who want to specialize in machine learning tend to go into computer vision labs. I was always curious about why computer vision in particular. My confusion was ...
1
vote
1answer
53 views

Text analysis : What after term-document matrix?

I am trying to build predictive models from text data. I built document-term matrix from the text data (unigram and bigram) and built different types of models on that (like svm, random forest, ...
3
votes
1answer
40 views

Predict user behaviour with constantly changing input variables

How to work on building an engine for a website wherein we want to score/recommend stuff based on her different activities, like the music she rated or the article she read, or whether email ...
0
votes
0answers
16 views

R: How to choose the height parameter in cutree, or: how to find the optimal number of clusters in UPGMA clustering?

I am using hclust() to carry out a UPGMA clustering (method="average") in R. Then, I'm using ...
0
votes
0answers
18 views

Rank Deficiency?

What can I do about rank deficiency? Can I just ignore it? I have an unbalanced dataset and using logistic regression (caret glm). I get 50 errors saying that my data is rank deficient and the results ...
0
votes
0answers
13 views

Search terms and Semantics [closed]

Suppose I type apple and get the following list of results: ...
0
votes
0answers
17 views

Describing the distribution of N points in D-dimensional space?

I want to tackle a classification problem by describing the samples as its descriptors' distributions. So let's say each sample has a label, and $N$ vectors of dimension $D$, (N and D are fixed) and ...
2
votes
1answer
82 views

How do you use test data set after Cross-validation

In some lectures and tutorials they suggest to split your data in three parts: training, validation and test. But it is not clear how test data set should be used and how this approach is better than ...
11
votes
3answers
277 views

How to intuitively explain what a kernel is?

Many machine learning classifiers (e.g. support vector machines) allow one to specify a kernel. What would be an intuitive way of explaining what a kernel is? One aspect I have been thinking of is ...
0
votes
0answers
18 views

machine learning for a ontology classification problem

I am working on a ontology based classification problem.The main objective was: computing ontology has keywords related to different categories.Each category talks about the domain it is related.For ...
1
vote
0answers
29 views

Limitations of ensemble selection from libraries

Question related to the approach in Caruana's paper: "Ensemble Selection from Libraries of Models" (linked below) http://www.cs.cornell.edu/~caruana/ctp/ct.papers/caruana.icml04.icdm06long.pdf Seems ...
1
vote
0answers
19 views

CTR Enhancement Model

I am looking to build a model to enhance CTR. Below is the business description. We have a coupon based website. For each retailer, we have a retailer page. Each retailer page will have up-to 100 ...
0
votes
0answers
10 views

how to calculate the number of parameter (no zero parameter) in an matrix in matlab I give the model_matrix [closed]

fonction nmbrerparameter=model_matrix c=0 for i=1:10 for j=1:10 if c~0 size(model_matrix.trans) size(mode_matrix.mix) end end end
0
votes
0answers
28 views

Which unsupervised learning method should I use on classification on many point cloud datasets?

I have a few abstract and high dimensional point clouds in the form of distance matrices. I want to do unsupervised learning on this dataset. The problem is, I am not using one distance matrix, but ...
0
votes
0answers
18 views

Why is linear svm solver faster than nonlinear solver?

Both linear and non linear SVM solve can be solved using primal or dual problem. Why is linear svm solver faster than nonlinear solver?