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

learn more… | top users | synonyms (1)

0
votes
1answer
36 views

The performance of the classifier went bad after adding extra two features

I'm solving a classification problem using 10 features and logistic regression. The performance of the classifier is fine when I use the 10 features only, however when adding another 2 features, the ...
0
votes
0answers
14 views

Which classification techniques output a probability that a sample is positive?

For example, logistic regression can predict that a sample will have label 0 or 1. But if you ask R to do a logistic regression, it won't output 0 or 1, it will output a probability that the label is ...
0
votes
0answers
18 views

Am I fine in implementing Naive Bayes Classifier?

I have implemented one Sentiment Analysis using Naive Bayes Classifier. To do this I have taken the following steps. First I checked the problem, the nature of data (continuous or discrete) and ...
0
votes
1answer
22 views

Minimizing the misclassification rate

I am reading the book Pattern Recognition and Machine Learning, and have a specific question from a text snippet. I'll state a few lines in the text Suppose that our goal is simply to make as few ...
3
votes
0answers
65 views

Need pointers to deep learning tutorials

I'm looking for good study material about deep belief networks, with particular emphasis to classification and feature extraction tasks for non-image data. I don't seem to find a great deal about ...
4
votes
1answer
107 views

What does it mean to say that “a topic is a distribution on words”?

I was taking a machine learning course and they say the following two phrases that confuse me: each document is a distribution on topics. and each topic is a distribution on words. I was ...
0
votes
0answers
20 views

Can a labeled LDA (Latent Dirichlet Allocation) dataset have just one label per document?

I understand that in labeled LDA, every document should be associated with a set of labels which are known as tagged topics for the respective document. My question is whether a document can be ...
1
vote
0answers
32 views

predict sales using naive bayes and handle sparse data problem

Problem I am trying to use naive bayes for ranking products in a search application. I would like to predict the sales of a given product given the search keyword and the category. the current formula ...
1
vote
0answers
14 views

Whats is meant by proximity in random forests?

I came across the term proximity in random forest. But I couldnt understand what it does in random forest. How does it help in classification??
0
votes
0answers
30 views

Hessian-Free instead of LSTM for Recurrent Net Machine Translation

Last year, Ilya Sutskever and collaborators came out with a paper about a recurrent LSTM net that learns sequence to sequence mappings for machine translation. It's somewhat surprising that the ...
0
votes
1answer
25 views

How to optimize Gaussian-process parameters for multiple tasks with GPML?

I have a lot of test curves and I want to optimize the length and scale parameters simultaneously for all curves. Is this possible?
0
votes
0answers
24 views

Relationship between kernel function for distance (locally weighted regression) and kernel function for SVMs?

I am reading Tom Mitchell's Machine Learning. In section 8.2.3, he defines: Kernel function is the function of distance that is used to determine the weight of each training example. In other words, ...
1
vote
0answers
61 views

Hierarchical clustering, linkage methods and dynamic time warping

My goal is to cluster time series based on their DTW distance. Therefore I've calculated full distance matrices as input for several clustering algorithms. I first had a look at hierarchical methods, ...
4
votes
2answers
29 views

Picking cases to label for classification

I'm working on a classification problem with the goal of diagnosing kidney diseases from clinical data. For each patient, we have a large number of observations, and would like to determine whether a ...
0
votes
0answers
7 views

Sample complexity for agnostic PAC learning for real valued functions

How many samples are needed for ERM to have $\epsilon$ excess risk relative to the best hypthosis $h^*$? Assume a bounded (and Lipschitz, if needed) loss function. The only survey I have been able to ...
3
votes
1answer
74 views

What is the difference between feature selection and dimensionality reduction?

I know that both feature selection and dimensionality reduction aim towards reducing the number of features in the original set of features. What is the exact difference between the two if we are ...
1
vote
1answer
33 views

Does the recurrent neural network require the length of input samples all the same

Theoretically, the training of RNN doesn't require that the samples must have the same time length, but it seems to me that some software or open-source requires that the input data has the same time ...
0
votes
1answer
41 views

Least-square fit with uneven distribution of data

I'd like to perform least-squares fit to data which is unevenly distributed on the x-axis. For example, if I was to bin the data, it would be something like x = 0~5: 10 data points x = 5~10: 20 ...
1
vote
1answer
25 views

Correlated cases and Cross Validation

I'm posting to ask if there is a method of cross-validation for correllated data that is already well implemented in R language. Some quick search on such method shows some techniques like h-block ...
1
vote
0answers
34 views

What is the advantage of non-negativity in matrix factorization?

I am wondering why matrix factorization techniques in the machine learning domain almost always expect the provided matrix to be non-negative. What is the advantage of this constraint? Background: I ...
0
votes
0answers
31 views

How do I determine the minimum sample size needed to evaluate the performance of a binary classifier with a statistically significant result?

I’m trying to determine the minimum sample size(s) needed to evaluate the performance of a binary classifier with a statistically significant result. I'm not trying to compare different classifiers ...
4
votes
3answers
103 views

What is the straightest line I can make using a linear combination of time series

I have 3 processes which generate an output in the form of a time series. I want to choose a linear combination of the processes that will result in the straightest line possible (I think this ...
1
vote
0answers
16 views

How to find the probalities of a,b,c (mutually exclusive events) when i turn the problem into three binary classification problems?

I am using scikits linear logisitc regression to classify three events a,b and c. it works better (score) when i convert them into a binary classification model. such as: 1. M1 classifies a or b 2. ...
0
votes
0answers
10 views

Personalization, user-adaptive and recommender systems

I am currently undergoing research into the field of systems that adapt content and layout depending on how the user uses the application. I am however puzzled as to the following terms as they are ...
0
votes
2answers
61 views

Data partitioning according two variables

I am working with the following dataset: ...
2
votes
2answers
33 views

Should I use epochs > 1 when training data is unlimited?

If I have virtually endless training data (it's synthesized) is there still purpose in having epochs? I.e. training on the same samples multiple times?
1
vote
1answer
64 views

Popular named entity resolution software

I am working on a project and need to extract persons' names from a large amount of documents. This task should belong to the named entity resolution problem. What are currently some of the most ...
0
votes
0answers
7 views

Estimating the covariance matrix in LDA ESL

I am reading Elements of Statistical Learning on LDA. On page 109, it talks about how we need to estimate parameters for Gaussian distribution. But why do we use this estimate for the covariance ...
0
votes
1answer
37 views

Concept of Hopfield Networks

Just want to double check my understanding of comcept of Hopfield networks. would a trained hopfield network have an energy function that has local minimas equal to the number of the training ...
2
votes
2answers
107 views

How to understand that MLE of Variance is biased in a Gaussian distribution?

I'm reading PRML and I don't understand the picture. Could you please give some hints to understand the picture and why the MLE of variance in a Gaussian distribution is biased? formula 1.55: $$ ...
2
votes
2answers
61 views

Odds vs. Probabilities (Confusion)

In logistic regression, I know that Odds and Probabilities are two different things, but still I can't distinguish the difference between them. I will be very grateful with your clarifications.
0
votes
0answers
24 views

Handling sparse document term matrix

I am working with a corpus of several thousand documents (41,732) however the documents tend to be short (the median number of terms per document is 3) resulting in a sparse document term matrix. ...
0
votes
0answers
20 views

To interpret SVM's probability output in text analysis

So, my question is about the application of the SVM's (or naive Bayes') probability output (via Platt's scaling). I know the interpretation of the output is the probability of a given observation ...
0
votes
0answers
18 views

How to Format Data for Structured Learning Problem?

I'm working on a project classifying discussion forum posts into various pre-defined categories, and would like to use a sequential learning model such as CRF's. I code mostly in Python and have found ...
0
votes
0answers
39 views

analysis of bank account record

I am new to the field of time series analysis, but I would like to have a look at my bank account and determine my spending habits. I read a lot about clustering of multiple time series but I think I ...
0
votes
0answers
43 views

What do these decision boundaries indicate in random forest and svm?

I was working on data science harvard homework problem. It is a two class classification problem in which they plot the decision boundary for random forest, svm and decision tree. The problem has 2 ...
1
vote
1answer
52 views

Are visualization techniques useful when the predictive model is a highly flexible machine learning algorithm?

I’ve always used highly flexible machine learning algorithms like boosted trees, support vector machines, and random forests that supposedly excel at identifying non-linear and irregular patterns and ...
1
vote
1answer
35 views

Understanding the effect of hyperparameters in machine learning experiments

In machine learning every algorithm has a set of hyperparameters which needs to be optimized for best prediction performance. The simplest method for this optimization is called grid search which ...
1
vote
1answer
47 views

Conditional distribution for Exponential family

We have a random variable $X$ that belongs to the exponential family with p.d.f. $$ P_X(x|\boldsymbol \theta) = h(x) \exp\left(\eta({\boldsymbol \theta}) . T(x) - A({\boldsymbol \theta}) \right) $$ ...
0
votes
1answer
49 views

determining how “important” a feature is in predicting a target in decision trees

Random forests allow us to compute a heuristic for determining how "important" a feature is in predicting a target. This heuristic measures the change in prediction accuracy if we take a given ...
1
vote
0answers
36 views

In developing a personal assistant, what is the best classification machine learning algorithm to use? [closed]

I've been developing a personal assistant-type program (siri-esque, I guess), which currently uses regexes loosely matching text to infer the requested operation, but as you can imagine, as it grows, ...
0
votes
0answers
48 views

Any machine Learning models to predict dates?

I have a general question regarding machine learning models. The idea is to predict what DATE the customer is likely to make transactions or purchases. Variables present in the data set are item, ...
1
vote
0answers
17 views

Early electric perceptron [closed]

As I was reading up about the perceptron, wikipedia has a tidbit that says an early prototype of the perceptron had potentiometers for weights and servos to change the weights. Would any one have a ...
0
votes
1answer
71 views

Leave-One-Subject-Out cv method

I would like to use a Leave-One-Subject-Out cv on my datasets (I have dataset including 38, 15, 10 participants, respectively). I don't know the hyperparamenters C and gamma of my SVM so I have to ...
2
votes
1answer
80 views

How to handle data normalization in kNN when new test data is received

I had a discussion with my colleagues about the following problem: Lets say we have 100 points of labeled data and we are using $k$-nearest neighbor method for prediction. So our data looks like ...
6
votes
1answer
69 views

VC-Dimension of k-nearest neighbor

What is the VC-Dimension of the k-nearest neighbor algorithm if k is equal to the number of training points used? Context: This question was asked in a course I take and the answer given there was ...
0
votes
0answers
19 views

Multiple Transformations of a Dependent Variable

Assume you have a data set with a continuous output variable (y), five continuous predictors (x1, x2, x3, x4, and x5), and a few additional categorical predictors. x1 through x4 have a weak ...
0
votes
1answer
23 views

Is the solution of SVM classifier a vector in second conjugate space of the RKHS

Let the training points be given by $x_1, x_2 \cdots x_m$. Suppose we want to predict the class of a new point $x$ as $f(x)$. In an linear SVM this is a dot product usually denoted by $<w,x>$ ...
3
votes
0answers
52 views

State of the Art versions of Generalized Additive Models

Generalized Additive Models [Tribshirani 86] was well received with over 1335 Citations. I am also aware of the popular version of GAM - the Multivariate Adaptive Regression Splines [MARS by Friedman ...
6
votes
4answers
492 views

Difference between Bias and Error?

In statistics, what is the difference between Bias and Error? You can say, Bias is a type of error? or Bias is an error with some tendency?