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

learn more… | top users | synonyms (1)

0
votes
0answers
15 views

Optimal Margin Classifer : Optimization Problem Setup

In the notes from Andrew Ng Machine Learning course, he writes the initial optimization problem as follows. I am confused by the notation and suspect I am missing something simple. Given the ...
0
votes
0answers
26 views

Greedy subtree selection in Nested Hierarchical Dirichlet Processes

I'm implementing the Nested Hierarchical Dirichlet Process as described in this paper by Paisly et. al, 2014: http://arxiv.org/abs/1210.6738 My question is about the variational objective in Equation ...
0
votes
1answer
22 views

What is the intuition behind the Kappa statistical value in classification

I understand the formula behind the Kappa statistic value and how to calculate the O and E value from a confusion matrix. My question is what is the intuition behind this measure? Why does it work so ...
0
votes
1answer
24 views

How to choose an appropriate maxdepth in rpart.conrol?

I'm using the boosting method in adabag library and trying to choose an appropriate maxdepth in rpart.control for building a 2-class classification model using my training dataset. I have noticed that ...
0
votes
0answers
16 views

Can you use accelerometer data for classification with Conditional Random Fields?

I want to recognize activities, based on accelerometer data from the smartphone. I studied Conditional Random Fields and the CRFSuite. Now I am Confused. In my opinion CRF training uses static single ...
0
votes
1answer
38 views

Vector Space Model for Online News Clustering

I am trying to automatically cluster news articles based on their content. I need this algorithm to be online and simply group news articles related to the same story as they arrive. The common ...
1
vote
0answers
15 views

Classification with two different dataset

I am working on a cancer classification model.Task is ,I am initially given a data set of 500 people and 1000 features.These people are given some kind of treatment(say Treatment 1). Some people are ...
0
votes
2answers
68 views

Split clustered data into calibration and validation sample (Cross validation)

I have a dataset with >800 cases ($n$) from >30 ($k$) different organisations (clustered data). The number of cases within each organisation differ (unbalanced data; e.g.: organisation 1 = 30 cases, ...
2
votes
1answer
65 views

R caret package - number of principal components when preprocessing using PCA

I am using the caret package in R for training of binary SVM classifiers. For reduction of features I am preprocessing with PCA using the built in feature [preProc=c("pca")] when calling train(). How ...
1
vote
1answer
73 views

Better in ROC AUC vs. better in PR AUC

I'm comparing two classification models by computing the area under ROC and Precision-Recall curves. However sometimes one model is better with AU-ROC but worse in AU-PR, and other times it's better ...
0
votes
1answer
75 views

How does Support Vector Machine compare to Logistic Regression?

Support Vector Machine (SVM) and logistic regression (LR) have been discussed widely in machine learning community, I know that both of them achieve pretty good performance. But, I am not sure how in ...
0
votes
0answers
19 views

ANOVA and Principal Component Regression

I Just need your valuable suggestions. ...
1
vote
1answer
36 views

Data Augmentation using Eigenvalues and Eigenvectors

Recently, I have come across a paper which has used a unique way of augmenting the data. If the data has multiple channels say we have a $x_i , i=1...N$ as a column feature vector. If we compute the ...
1
vote
2answers
44 views

How to best to use Continuous value features with discreet values for logistic regression based binary classification problem

This is related to Minimisation algorithm for a mix of discreet and continuous parameters? I am trying out logistic regression to solve a binary classification problem. Though I am feature-scaling ...
1
vote
1answer
35 views

Understanding SVM and when to precompute the normal vector

I've been reading a lot about SVMs and have some questions about performing classification from the SVM model produced from a package like libSVM. From my understanding, for a linear SVM without the ...
0
votes
0answers
35 views

One or two output neurons for a binary classification task with an artificial neural network

Suppose you have a classification problem in which you want to classify inputs into two exclusive classes (y1 and y2) with an artificial neural network (which models P(y|x)). Among the two following ...
3
votes
1answer
62 views

“…if the data is linearly separable”

I keep hearing this phrase as a precursor to many algorithms, but I am not sure how exactly one goes about finding out if the data is indeed, linearly separable. Of course, if the data has ...
2
votes
2answers
55 views

Using a gaussian kernel in SVM. How exactly is this then written as a dot product?

I am attempting to use SVMs for my class project. For this project, I have selected the gaussian kernel as, well, the kernel. That is, $$ k(\mathbf{x}_1, \mathbf{x}_n) = e^{-\gamma ||\mathbf{x}_1 - ...
0
votes
0answers
32 views

R PNN slow- other packages?

I'm trying to run a pnn (Probabilistic neural network) in R, using pnn package. It's function smooth uses rgenoud package to optimize. I only used 9k lines and it's real slow. Is there any other ...
0
votes
1answer
39 views

What is AUC of PR-curve?

I understand that AUC under ROC curve is a classic evaluation measurement for classifiers (which is basically the accuracy). However, when data is imbalanced, PR will be alternative. So, what does the ...
0
votes
1answer
32 views

Pattern recognition in state sequences

I have a sequence of states of a system. Each state is defined by an abstract identifier e.g "Eating", "Sleeping" etc... and a duration. So a state is basically ...
0
votes
1answer
28 views

Have difficulty understanding Matlab's Ridge regression

I am confused by Matlab's documentation of Ridge regression at http://www.mathworks.com/help/stats/ridge-regression.html and couldn't figure it out by myself. On that page, the Introduction to Ridge ...
0
votes
0answers
30 views

SVM One-vs-One vs One-vs-ALL SVM

For an unbalanced dataset annotated by human annotators in which each item is assigned to different classes, what is the argument for and against using any of One-vs-One vs One-vs-ALL SVM ...
2
votes
1answer
82 views

How to determine the optimal threshold for a classifier and generate ROC curve?

Let say we have a SVM classifier, how do we generate ROC curve? (Like theoretically) (because we are generate TPR and FPR with each of the threshold). And how do we determine the optimal threshold for ...
1
vote
2answers
46 views

Strategies for parallelising neural networks

When it comes to parallelising a problem, it involves the division of routines and subroutines between a number of nodes, namely; the master node and the slave nodes. Once each of these nodes ...
1
vote
1answer
54 views

Predicting continuous output

I'm trying to predict output per worker for given inputs of capital (physical capital), labor (human capital) & productivity. I have a data set of several countries ...
0
votes
2answers
69 views

What are the differences between AUC and F1-score?

F1-score is the harmonic mean of precision and recall. The y-axis of recall is true positive rate (which is also recall). So, sometime classifiers can have low recall but very high AUC, what that ...
4
votes
0answers
71 views

Rademacher complexity of logistic regression

Consider logistic regression. We have the logistic loss function, $\phi: R\rightarrow [0,1], \phi(u)=\log(1+\exp(-u))$, which is Lipschitz, and we have the linear function class $F=\{f_w:R^d ...
1
vote
2answers
85 views

How to know which features in a linear classifier mainly led to a prediction?

I have a classification problem where I use a model (say Logistic regression or SVM) to determine whether an instance belongs to class 0 or class 1. For a certain prediction on a test instance X, if ...
0
votes
0answers
20 views

Create a damping function for discrete time series data such that values converge to constant value

I have an agent-based model where an agent predicts output and then compares that value to the actual output. How can I create a damping function of sorts that will cause the delta between expected ...
0
votes
0answers
30 views

My support vectors don't look correct

I am trying to classify a toy dataset using SVM. I only have two features and 20 instances. The decision boundary seems correct, however, the support vectors dont look correct. This is the relevant ...
0
votes
0answers
43 views

Unsupervised learning

I am implementing a paper titled as "Internet Traffic Identification using Machine Learning" by J. Erman et al.. I have completed supervised part, but now I got stuck in unsupervised learning part. ...
0
votes
0answers
22 views

Positive and negative examples in Rocchio-based recommender

I am exploring the usage of Rocchio-based recommenders in e-commerce and news portals and trying to wrap my head about the concept of a negative rating. Often in e-commerce or news portals there is no ...
0
votes
0answers
26 views

Why we even try to minimize a loss function, which is non-convex, in matrix factorization?

In matrix factorization (especially under the scenario of recommendation system), we often try to factorize a matrix Y into two low rank matrices: $Y=U\cdot V^T$ If we assume there $m$ instances in ...
0
votes
0answers
31 views

How to increase Accuracy in Age model on Mobile operator dataset (Call log)?

I know question is too lengthy, but it need description. I am currently working on predictive analysis project. Which predict customer age, gender, etc. demographics attributes from their data set. ...
1
vote
0answers
16 views

Active Learning for Tweets Classification

I am working on a binary classification problem on Twitter, in which I have to classify tweets in one of two categories. Now I would like to test how different active learning frameworks would work ...
0
votes
1answer
39 views

Is it possible to combine several clustering results in a meaningful way?

The problem I face is somewhat awkward, I have 40,000 points in my dataset and I would like to cluster them hierarchically. But due to the limitation of my laptop(and R) in each run of clustering only ...
0
votes
0answers
21 views

How do I represent semi-supervised learning in plate notation?

In a plate notation, how do I represent variables in which both labeled and unlabeled instances exist? For example, let's say I have an Latent Dirichlet allocation-like algorithm in which some ...
1
vote
2answers
76 views

Help with understanding statistical measures and Receiver Operating Characteristics

In my machine learning class we just went over statistical measures and plots. We looked at the definitions of True Positive Rate (sensitivity/recall, etc), 1-False Positive Rate (speciicity), ...
13
votes
5answers
1k views

Why downsample?

Suppose I want to learn a classifier that predicts if an email is spam. And suppose only 1% of emails are spam. The easiest thing to do would be to learn the trivial classifier that says none of the ...
0
votes
1answer
56 views

How to solve this problem on Curse of Dimensionality problem - Nearest Neighbours

I have started learning classification techniques and trying to solve the problems from the book Introduction to Statistical Learning. While currently working on the which is based on Curse of ...
0
votes
0answers
20 views

How do I implement custom sparse connections in a neural network?

We'd like to implement neural network which is not fully connected - we want to explicitly set which output neurons connect to which input neurons (there's no hidden layer). We use Theano but we can ...
1
vote
0answers
12 views

Applying linear function approximation to reinforcement learning

How do you apply a linear function approximation algorithm to a reinforcement learning problem that needs to recommend an action A in a specific state S? I've read over a few sources, including this ...
2
votes
0answers
38 views

How to compare two ranking algorithms?

I want to compare two ranking algorithms. In these algorithms, client specifies some conditions in his/her search. According to the client`s requirements, these algorithm should assign a score for ...
1
vote
2answers
75 views

How to do cross-validation when comparing different feature selection methods?

I am using SVM for a prediction task. My sample size is small, only N=140. Suppose I want to compare the prediction accuracy when using two different feature selection methods. Would it be better to: ...
0
votes
1answer
60 views

When normalization is counter-productive [duplicate]

Could you give me general examples of when normalization is not used properly and affects badly the classification accuracy, or when it is not needed?
0
votes
0answers
11 views

Event level driven response modeling

I am investigating operational and maintenance data for a fielded system. There is a year worth of data. The operational data has been reduced to fault indications, which are triggered when ...
2
votes
2answers
82 views

Does Support Vector Machine handle imbalanced Dataset?

Does SVM handles imbalanced dataset? Is that any parameters (like C, or misclassification cost) handling the imbalanced dataset?
0
votes
0answers
35 views

difference between machine learning and stastitical technique [duplicate]

Is there any difference between machine learning and stastitical techniques. I have searched a lot some researchers say that there are some overlap some are saying there is no difference.Can you give ...
1
vote
0answers
18 views

Assessing predictor contribution to model output

Many of machine learning methods are considered as "black boxes". Examples of such methods are SVM, Neural Networks, Random forests etc. One may apply sensitivity analysis techniques (as described for ...