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

learn more… | top users | synonyms (1)

1
vote
1answer
17 views

Is it possible for test error to be lower than training error

Is it possible to have test error lower than training error? I have a classification problem with 2000 samples, 500 of which are positives, 1500 are negatives. I split my data into 70% training data, ...
1
vote
1answer
39 views

Use of a bagging model or feature engineering?

As a pet project, I have been learning some data analysis and machine learning skills (mainly text analytics) with the Analytics Edge course on edX. I decided to put some of my new skills at use ...
0
votes
0answers
9 views

Representative signal

I'm implementing machine learning with sensor data. I am having the problem that some sensors not always have good integrity, that is, not all data points arrive at destination because of ...
0
votes
1answer
17 views

Notation - stacked vectors

I am trying to self-study Kevin Murphy's book on machine learning. i am trying to be 100% sure that i understand the notation in the book. . I am however struggling to understand what is meant with ...
0
votes
0answers
34 views

How does the convolution work for a simple example 1D and its relation to the true mathematical convolution?

I was trying to pin point precisely mathematically what the convolution does for a simple 1D example (i.e. $x \in \mathbb{R}^D$ as opposed to an image $x \in \mathbb{R}^{D_1 \times D_2}$). The ...
1
vote
1answer
47 views

A question about Dynamic Random Forest

On this article, Simon Bernard proposes a new approach for constructing Random Forest called Dynamic Random Forest. I am new on this subject, so after reading the article, I have a doubt regarding the ...
0
votes
0answers
19 views

Integrate a function over variable following normal distribution in R [migrated]

I want to integrate the inverse logit function over a variable $X$, which follows a normal distribution, $X\sim N(21,7)$. I'm a beginner in R so i have been trying following commands that didn't work ...
2
votes
0answers
79 views

Convolutional Neural Networks with Caffe and NEGATIVE IMAGES

When training a set of classes (let's say #clases (number of classes) = N) on Caffe Deep Learning (or any CNN framework) and I make a query to the caffemodel, I get a % of probability of that image ...
2
votes
1answer
41 views

What is the point of graphical models?

I spent the day learning about the bnlearn package in R only to discover that Bayesian models do not work with undirected graphs. I'm trying to learn about the Markov Random Field Network, and so far ...
0
votes
0answers
32 views

Sampling, feature selection and preprocessing in cross validation

To brief my question, I want to clarify the order of parameter tuning and the correctness of the flow in my scheme. In my classification scheme, there are several steps including: SMOTE (Synthetic ...
1
vote
0answers
10 views

Given a data table with outlier data points, how to determine whether the data table include global outliers or local outliers?

As we know, outliers are categorized into (i) global outliers and (ii) local outliers. Local outliers are outliers comparing to their local neighborhoods, instead of the global data distribution. ...
2
votes
1answer
79 views

Why time series analysis is not considered a machine learning algorithm

Why is time series analysis not considered a machine learning algorithm ( unlike linear regression). Both regression and time series analysis are forecasting methods. So why is one of them considered ...
4
votes
1answer
57 views

How can 1 more feature disrupt a Random Forest's confusion matrix?

I'm trying to predict a binary variable with both random forests and logistic regression. I've got unbalanced classes (approx 1.5% of Y=1), so i'm calling ...
0
votes
1answer
37 views

How does svm deal with new levels of a variable added over time when considering time series data?

I am trying to predict customer spending for an X year period after t0. I train an svm model with transactions occurring before and on t0, on the cumulative spending of the customers after t0. I then ...
1
vote
0answers
30 views

Temporal Difference Learning Not Converging

I'm trying to implement a $Q(\lambda)$ algorithm from this paper (warning: link is a download of a PDF) and can't seem to get it to find anything that close to the optimal policy. From what I can ...
0
votes
1answer
30 views

Does a theoretical “perfect (accuracy) score” exists we could target for a given dataset?

My question is the following : You have a dataset, and you want to determine theoretically what accuracy score (or other way to measure performance such as AUC, etc.) a "perfect" model could get on ...
0
votes
0answers
6 views

Use known output distribution to increase maxent classifier performance

If you know the distribution of your output, can you use this information to improve the performance of a maximum entropy classifier? e.g., I know that a gallery of pictures is 40% cats, 60% ...
1
vote
0answers
49 views

Regression models to only predict integers (instead of floating point numbers)?

I have a dataset that consists of about 50 different attributes. One of these attributes I want to predict, using the other attributes as features. The values of the attribute that I want to predict ...
1
vote
2answers
59 views

Do we need gradient descent to find the coefficients of a linear regression model

I was trying to learn machine learning using the coursera material Andrew Ng uses gradient descent algorithm to find the coefficients of the linear regression model that will minimize the error ...
6
votes
4answers
1k views

Is automated machine learning a dream?

As I discover machine learning I see different interesting techniques such as: automatically tune algorithms with techniques such as grid search, get more ...
0
votes
0answers
26 views

How do I use Hidden Markov Model Viterbi algorithm for sequence labeling?

with my current small experience of HMM. Given that i have some patterns (sequence of interest for example gestures or words in spoken language) if i need to use HMM for sequence classification ...
0
votes
1answer
16 views

Unconnected Linearly Seperable Classification

Consider classifying something like the case shown below (exagerated syntetic example): If this were a task to classsify into 3 groups, (blue-left, red, blue-right), then a Linear Support Vector ...
0
votes
0answers
8 views

Structural Correspondence Learning for Domain Adaptation - how is the augmented data formed?

Structural Correspondence Learning (SCL) is a method for dealing with domain adaptation (different data distributions in the training and testing sets). It was proposed by Blitzer et al. but I am ...
0
votes
2answers
55 views

Are fixed bias neurons or biased neurons better?

When building an artificial neural network, there seems to be two differing philosophies in usage of biases. There are those groups that propose neural networks with a fixed bias neuron with a ...
0
votes
0answers
12 views

Softmax Regression GD Update Derivation

I'm implementing softmax regression and am deriving the max-log-likelihood update for gradient descent by hand first. Coming from the Stanford UFLDL site, they show the gradient of the cost function ...
0
votes
1answer
16 views

how to interpret metrics for my model

I understand how these metrics are determined but can anyone give me typical poor, acceptable, good, and outstanding values RAE and RSE?
0
votes
1answer
42 views

How to do one-vs-one classification for logistic regression?

I have a dataset with 4 clases and I want to apply logistic regression with one-vs-one classification. So, first I train for each pair of classes a logistic regression classifier (i.e. calculate the ...
1
vote
1answer
40 views

Adaboost for numeric dataset

I have been trying to fit Adaboost to work with continuous valued data set and the more I read the more I keep getting confused. I have read about the multiclass Adaboost with log(K-1) addition to ...
0
votes
0answers
25 views

How can you detect outliers in a group of face images?

I'm trying to filter an image database which contains some irrelevant pictures. All the faces are labeled with points around the face contour, eyes, mouth, eyebrows, have age and gender. The faces are ...
0
votes
0answers
13 views

what exactly is projected gradient descent?

I am reading the article with title "metric learning by collapsing classes" http://papers.nips.cc/paper/2947-metric-learning-by-collapsing-classes.pdf.There are some questions which has bothered me ...
6
votes
1answer
150 views

What are the differences between Ridge regression using R's glmnet and Python's scikit-learn?

I am going through the LAB section §6.6 on Ridge Regression/Lasso in the book 'An Introduction to Statistical Learning with Applications in R' by James, Witten, Hastie, Tibshirani (2013). More ...
1
vote
1answer
24 views

Data splitting using kennard stone [closed]

I would like to use kennard stone algorithm for splitting a dataset into training and test set. Does Weka /Rapidminer/any free software with GUI has a feature to do this? If so, would like to know the ...
1
vote
1answer
39 views

Why is the energy and the probability of a configuration related in a Boltzmann machine?

According to Hinton's slides (slide 34) the following relationship holds for Boltzmann Machines in thermal equilibrium: $$ p(v,h) = \frac{e^{-E(v,h)}}{\sum_{u,g}{e^{-E(u,g)}}} $$ To me this is not ...
0
votes
0answers
37 views

How is bootstrapping used for machine learning?

How does one use bootstrapping in a machine learning context? My typical data analysis pipeline is Split data into 10 folds Train classifier with 9 folds Test classifier with remaining fold Repeat ...
0
votes
0answers
23 views

Get important features of n samples

Suppose I have a data frame of [n_samples, m_features] with the corresponding variances of the features [n_features]. The values in my data is between 0 and 1 so the question is: Is there any way to ...
0
votes
0answers
33 views

How can I make Weka classify the smaller class, with a 2:1 class imbalance?

How can I make Weka classify the smaller classification? I have a data set where the positive classification is 35% of the data set and the negative classification is 65% of the data set. I want Weka ...
0
votes
0answers
27 views

Time-series cross-sectional classification problem

I have a time-series cross-sectional dataset consisting of 100 individuals that each had 4 features measured yearly for 21 consecutive years. One of the features is binary and the other three are ...
0
votes
0answers
25 views

Representation of misspelled words for neural network?

While thinking about a neural network based spellchecker, I was thinking about word embedding not being able to represent any "unique" (misspelled) words that the model haven't seen before. I tried ...
1
vote
0answers
9 views

how can fixed parameters cost and gamma using libsvm matlab to improve accuracy? [migrated]

I use libsvm to classify a data base that contain 1000 labels. I am new in libsvm and I found a problem to choose the parameters c and g to improve performance. First, here is the program that I use ...
0
votes
0answers
37 views

Kernel K nearest neighbours with sparse data

I have a big sparse matrix (around 5 million of lines, 20 000 predictors), and I would like to run a kernelized k-NN on it. However, I don't know how to scale the data properly. So far, I have scaled ...
0
votes
1answer
69 views

Improving a logistic regression model in R [closed]

I have to devise a model in R, capable of predicting the type of a disease, which is a categorical dependent variable (three possible values), through several continuous and categorical, some of the ...
0
votes
0answers
17 views

What techniques infer dependencies between time series?

I'm working on an ML project. We have an app where users can track 1) binary variables and 2) quantities over time on a scale from one to ten. For example, at a given time they may track whether they ...
0
votes
0answers
16 views

VC Dimension of the set of canonical hyperplanes

This is a proof of the theorem about VC Dimension of the set of canonical hyperplanes from Professor Mohri's lecture slide. I'm having difficulty with understanding the inequality$$ \forall i \in ...
0
votes
0answers
10 views

RSNNS neural networks, checking percentage correct.

For those who have some experience with RSNNS, I'm trying to build a neural network for reading aloud, using RSNNS in R. To give some information about what I'm doing and using. I'm using orthographic ...
1
vote
1answer
20 views

Is there a difference between on-line learning, incremental learning and sequential learning?

What I mean is the following: Instead of processing all the training data at once and calculating a model or hypothesis, we process one data point at a time and update the model directly afterwards. ...
0
votes
0answers
10 views

How should I measure difference between user behavior / model performance on different population?

I am developing a recommendation engine whose goal is to suggest data exploration routes to non-technical users. The underlying model is content based, with the training data made up of the behavior ...
0
votes
0answers
21 views

How to use Particle Swarm Optimization for finding optimal bandwidth with cross-validation?

I want to use Particle Swarm Optimization (PSO)for finding optimal smoothing parameters of a kernel density estimation problem. Initially I tried to find the same using grid search method,but the ...
1
vote
0answers
10 views

Touch times to authenticate user

for a project I gathered touch data of different users when they tap a rhythm repeatedly on the touch screen in a game. ...
0
votes
0answers
14 views

Scikit Random forest pred_proba gives rounded off values

I am using random forest in scikit learn for classification and for getting the class probabilities , I used pred_proba function. But strangely it outputs probabilities rounded to first decimal place ...
0
votes
2answers
34 views

Name some techniques similar to Random Forests

I'm interested in what techniques are out there that are similar to, but not the same as, Random Forests. Either for classification or regression or both. Particularly interested in techniques which ...