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

learn more… | top users | synonyms (1)

0
votes
0answers
17 views

How to do sentence boundary detection? [closed]

I have a string as My name is sushil feeling bad My age is 30 and hobbies are none thanks for asking me name varun i am good at cricket and my age is 40 i have done a lot of work in research area , ...
0
votes
0answers
7 views

Comparing the performance of two classifiers for statistical significant differences

I am aware that there are similar questions asked on Cross Validated, however, this differs slightly. Say that I have one dataset, which I want to use to train two classifiers A and B (supervised ...
0
votes
0answers
26 views

How to deal with categorical target variable that has more categories in prediction than training?

I'm building a logistic regression model and found out that with my categorial target variable there are more categories in my prediction set than my training set. To be clearer: In e.g. my training ...
2
votes
0answers
25 views

Why doesn't the hidden state of a neuron network provide better dimension reduction result than original input?

I just read a great post here. I am curious about content of "An example with images" in that post. If the hidden states mean a lot of features of the original picture and getting closer to final ...
0
votes
0answers
9 views

In simple terms, what is a radial basis machine (RBM) modeling?

I stumbled on this term while trying to read this research article: http://clincancerres.aacrjournals.org/content/21/1/175 I don't have a background in statistics, so simple terms would be nice. Also,...
0
votes
1answer
16 views

Twitter Classification of tweets related to Ebola into 21 custom categories

I have a lot of twitter data (4GB) related to keyword Ebola. I want to classify the tweets into 21 categories. Categories :- Death - tweet is about death Health Care Workers - tweet is about ...
4
votes
2answers
44 views

Recommendation using Random Forest [closed]

I am trying to write my own recommender system. I have data set of user-item rating matrix. But I do not have profile information about either items or users. I already built pure CF using cosine and ...
0
votes
0answers
23 views

Classification(Machine Learning) or Survival Analysis

I am working on building prediction model for disk failures (time taken to occur a disk failure and what parameters could strongly affect disk failures). I am bit confused on- What data ...
0
votes
0answers
23 views

Product suggestions based on what users have also bought

We need to train a model that when given an input of a product name and group, it outputs product suggestions for what users frequently buy together. We have product data stored like: ...
0
votes
0answers
7 views

How to create machine learning model for sparse 1D data?

I want to predict the frequency of Dataset 1 from the information in Dataset 2. Dataset 1: List of 1D points on a flat number line. Number line ranges from 1 to 3 billion (yes billion). Dataset 2: ...
0
votes
0answers
26 views

Bayesian Information Criterion(BIC) for gaussian Mixture Models

Say that i have two classes, A and B and some training data $(x_i, y_i), y_i \in {A, B}$. My goal is to fit a mixture model into each class individually and calculate the BIC for the model. The BIC ...
0
votes
0answers
9 views

Is there any Revolution R functions ( Rx) functions available for Survival Random Forest Analysis? [closed]

I have R code which does Survival Analysis using Random Forests. It was taking huge amount of time. Is there any way to parallelize the operations using either Rev R Rx functions or any other way ?
3
votes
1answer
38 views

Regularized linear model: adding special constraints to the coefficient

I understand we can add $L_1$ or $L_2$ regularization to linear regression (Lasso and Ridge regression). In addition, it is possible to restrict the coefficient to be integers (see this post). ...
1
vote
0answers
14 views

time series classification of an event either happening or not happening using machine learning techniques

I have sensor data that I would like to use to classify whether an an event (giving birth) is about to occur within (2-4hrs) in an animal based on various metrics collected by the sensor(activity ...
0
votes
0answers
14 views

How to treat training example with multiple values for a feature

Problem Setting Suppose I am given a data set with the following: a user_id used to identify different users a list of books ...
0
votes
0answers
10 views

Can anyone help me understand the proof of the inequality that relates shatter number and Rademacher Complexity?

Let $F$ be a collections of binary function. $s(F, n)$ be the shattering number of $F$ defined as $sup_{z_1...z_n}|F_{z_1,...,z_n}|$. Can anyone explain the proof of the following inequality?
0
votes
1answer
23 views

Is there an optimal loss function for dealing with imbalanced classes?

I'm aware that there are many ways of dealing with datasets where there is a strong class imbalance in the target variable: downsampling the more prevalent and less important class, over-weighting the ...
0
votes
0answers
13 views

Bayesian Entropy Criterion

I am looking at the following paper which introduced the Bayesian Entropy Criterion as a way to select models that minimize the classification error rate: https://hal.inria.fr/inria-00070612/document ...
1
vote
3answers
53 views

Overfitting due to a unique identifier among features

In many beginner ML lectures / tutorials, it's advised to remove those features that uniquely identify the example. For example, if predicting user behavior, a numeric ...
0
votes
0answers
7 views

what should be validation parameter for Logistic Regression(LR) in online learning plus rare event scenario?

We have been following below paper to predict CTR( Click probability) of different ad items. This will be used to serve different ads based on probability values. http://olivier.chapelle.cc/pub/...
9
votes
0answers
91 views

Bridge penalty vs. Elastic Net regularization

Some penalty functions and approximations are well studied, such as the LASSO ($L_1$) and the Ridge ($L_2$) and how these compare in regression. I've been reading about the Bridge penalty, which is ...
-2
votes
1answer
52 views

If wrong retry - negating automatically incorrect answers [closed]

Say the first event is someone buying product, and the second them selling it. I need to predict how long they'll have the product for from today, but my dataset and my model only predict the overall ...
1
vote
0answers
16 views

Using Monte Carlo simulations with subsequent element removal

I'm attempting to build an evaluation set for a logistic regression classifier and I've run into a statistical problem. The study involves a very large population (G) that has two properties of ...
3
votes
1answer
23 views

Learning vector embeddings from distances

So... I have a set of entities $\mathcal{E} = \{e_i \mid i \in [1,n]\}$, and I have a proper distance metric defined over $\mathcal{E}\times\mathcal{E}$, call it $d$, so the distance between $e_i$ ...
4
votes
1answer
45 views

Does an optimally designed neural network contain zero “dead” ReLU neurons when trained?

In general should I retrain my neural network with fewer neurons so that it has fewer dead ReLU neurons? I've read conflicting opinions about dead ReLUs. Some sources say dead ReLUs are good because ...
0
votes
0answers
14 views

What is a good way to determine a major change has began in a set of numbers?

I'm visiting you guys from Stack Overflow because I need some non-programming related help with a project. So I am reading in a thermal image and trying to detect where the start of someone's head is. ...
13
votes
3answers
2k views

Can Machine Learning or Deep Learning algorithms be utilised to “improve” the sampling process of a MCMC technique?

Based on the little knowledge that I have on MCMC (Markov chain Monte Carlo) methods, I understand that sampling is a crucial part of the aforementioned technique. The most commonly used sampling ...
1
vote
2answers
44 views

Reinforcement learning algorithms dependency on the initial state

I would like to know whether in learning algorithms, convergence to the optimum depends on the initial parameters. I mean can you reach the optimum if you start from a given initial state, but you ...
1
vote
0answers
31 views

Why would a random forest model be biased towards sensitivity/specificity?

I am training a random forest model using the sk-learn library, for a binary classification task. For some reason, when I set the max_depth parameter to 1, the model has an average 90% accuracy on ...
0
votes
0answers
16 views

Why don't we do normalization for the output variable y?

In machine learning algorithms, we make the features to be in the same scale through feature normalization. Shouldn't the output variable of the training set be normalized as well, since the ...
3
votes
0answers
41 views

Can anyone explain the inner product within a Reproducing Kernel Hilbert Space?

I saw this formula on a machine learning paper, but just couldnt get my head around it. Can you anyone help me understand it? Suppose $H$ is a RKHS, K a kernel function, and $\{\lambda_1, \lambda_2, \...
0
votes
0answers
10 views

What an error rate in regression means ? [closed]

As I cannot say that the error in regression means precision as in classification. What an error rate after the system has been cross-validate in regression means ?
3
votes
2answers
33 views

Comparing clustering algorithms [duplicate]

I am conducting clustering analysis in which I am using three clustering algorithms K-means, Spectral Clustering, and ...
3
votes
1answer
42 views

Neural network library for Python for Microsoft Windows

I have been having trouble with selecting a good library for Neural network algorithms in Python. TensorFlow isn't supported on windows. Theano is still in the beta phase of development. PyBrain too ...
1
vote
1answer
18 views

Gaussian Processes Probabilistic Least Squares Classifier - Leave one Out Cross-Validation Means, Variances Shape?

I am currently working on implementing some of the algorithms covered in Rasmussen and Williams' book, and stuck on a particular part in chapter 5 (link to the chapter here). In particular, on page ...
2
votes
1answer
47 views

What does end to end mean in deep learning methods?

I want to know what it is, and how it is any different from ensembling? suppose, I want to achieve high accuracy in classification and segmentation, for a specific task, if I use different networks, ...
0
votes
1answer
23 views

How to embed prior distribution of datasets into the classification model

Given 3 training sets : $(X_1,y_1),(X_2,y_2)$ and $(X_3,y_3)$. These three datasets are separated as it is being manually tagged in the preprocessing. Based on the datasets, three classifiers can be ...
1
vote
1answer
34 views

Probability distribution estimate of target continuous variable

I am looking for litterature/reference on algorithms for a regression task that can give the probability distribution estimation of the output variable, or multiple outputs with their respective ...
0
votes
1answer
37 views

Request for reference of cheat-sheet of paper concerning preprocessing and learning algorithms [closed]

Is there a comprehensive list of preprocessing steps that are highly recommended when using the classical learning algorithms (see below for a list of the families)? Is there a cheat-sheet or a cook-...
0
votes
2answers
27 views

Why do Experimental Design when you could build classifiers?

Say that you wanted to test whether people who took a placebo were more likely to smoke. You can devise a experimental design/test for this, but I could also create a model that predicts whether ...
0
votes
0answers
16 views

Activations or Binary States for RBM Input

I understand that the activations or the sampled binary states can be used as input for an RBM or for the inputs of subsequent layers when pre-training a DBN - from the MILA DBN tutorial; Hintons 06 ...
0
votes
0answers
9 views

Comparing GMM Results on Different Matricies Representing the Same Documents

I'm working on a problem where I use bag of words and a Gaussian Mixture Model (GMM) to cluster documents. In addition to running GMM on the tf (term frequency matrix), I also try the following: ...
5
votes
2answers
134 views

By which ways can we, in principle, evaluate whether a model succeeded in generalizing?

A disclaimer: By using an informal term such as "generalize", I am aware I am getting close to philosophical territory, and that my question could be seen as unsuitable for CV. I will do my best ...
0
votes
0answers
8 views

Derivation of back propagation though time

There is a wonderful explanation for the implementation of Backpropagation through time in the this article by Denny Britz here: http://www.wildml.com/2015/10/recurrent-neural-networks-tutorial-part-...
1
vote
0answers
16 views

Determinant of confusion matrix: how useful is it as a performance metric?

I was thinking a bit about confusion matrices and it came to my mind the determinant of a confusion matrix could be an useful performance metric in classification. Indeed, I got some results in ...
0
votes
0answers
13 views

Is it the right to split a shuffled training set into training and validation sets?

I want to create a validation set for CIFAR10 dataset which can be found here. The training set has a file named train.txt which contains a list of image's path ...
0
votes
0answers
10 views

Is it accepted to fit model with standardized data and predict on non-standardized data? [duplicate]

If you standardize your training data, then can it work on unstandardized data during predictions accurately? Many algorithms require the feature data to be standardized and I am wondering how/why/...
2
votes
0answers
19 views

How are performance measures affected in PU learning?

When learning from only positive and unlabelled data (PU learning), how are performance measures affected, when compared to a standard supervised setting? For simplicity, let's assume that the entire ...
1
vote
1answer
18 views

Is it correct to transform SNP data (categorical) to numeric (0, 1, 2) format to apply ML algorithms later? Why not binary (0, 1) data?

I wanted to know why is it correct to transform SNP data to 0, 1, 2 format using a reference allele, for example: SNP1 with C/T alleles, transformation rules: CC = 2, CT = 1, TT = 0, to later apply ...
0
votes
1answer
21 views

SURF algorithm failing?

I am currently working on misalignment correction and after some research found a nice matlab toolbox for this purpose capable of doing Lucas-Kanade, ECC among others. I had success when using a ...