vonjd
  • Member for 11 years, 6 months
  • Last seen more than a week ago
What are i.i.d. random variables?
Accepted answer
79 votes

It means "Independent and identically distributed". A good example is a succession of throws of a fair coin: The coin has no memory, so all the throws are "independent". And every throw is 50:50 (...

View answer
Difference between confidence intervals and prediction intervals
52 votes

I found the following explanation helpful: Confidence intervals tell you about how well you have determined the mean. Assume that the data really are randomly sampled from a Gaussian distribution....

View answer
Data mining: How should I go about finding the functional form?
28 votes

To find the best fitting functional form (so called free-form or symbolic regression) for the data try this tool - to all of my knowledge this is the best one available (at least I am very excited ...

View answer
What's the difference between correlation and simple linear regression?
16 votes

All of the given answers so far provide important insights but it should not be forgotten that you can transform the parameters of one into the other: Regression: $y = mx + b$ Connection between ...

View answer
Why does the Lasso provide Variable Selection?
13 votes

I think there are excellent anwers already but just to add some intuition concerning the geometric interpretation: "The lasso performs $L1$ shrinkage, so that there are "corners" in the constraint, ...

View answer
Compute a cosine dissimilarity matrix in R
8 votes

You can use the cosine function from the lsa package: http://cran.r-project.org/web/packages/lsa

View answer
Proper way of using recurrent neural network for time series analysis
7 votes

Another possibility are Historical Consistent Neural Networks (HCNN). This architecture might be more appropriate for the above mentioned setup because they eliminate the often arbitrary distinction ...

View answer
Manually Calculating P value from t-value in t-test
7 votes

The best way to calculate it manually is: t.value = (mean(data) - 10) / (sd(data) / sqrt(length(data))) p.value = 2*pt(-abs(t.value), df=length(data)-1) You need the abs() function because ...

View answer
Why on average does each bootstrap sample contain roughly two thirds of observations?
5 votes

Just adding to @retsreg's answer this can also be demonstrated quite easily via numerical simulation in R: N <- 1e7 # number of instances and sample size bootstrap <- sample(c(1:N), N, replace =...

View answer
How to find true positive, true negative, false positive, false negative from a three class confusion matrix?
4 votes

When I understand your question correctly you are asking which class is the positive one and which is the negative one. The answer is that this is to a certain extent arbitrary, so you have to decide ...

View answer
Measuring non-linear dependence
3 votes

Please have a look at the following article from science - it addresses your point exactly: Detecting Novel Associations in Large Data Sets by David N. Reshef et al. From the abstract: ...

View answer
Why is accuracy not the best measure for assessing classification models?
2 votes

I wrote a whole blog post on the matter: https://blog.ephorie.de/zeror-the-simplest-possible-classifier-or-why-high-accuracy-can-be-misleading ZeroR, the simplest possible classifier, just takes the ...

View answer
Decision tree more than 2 output levels
2 votes

With the OneR package (which basically builds a one level tree with the best predictor) you can have any number of levels in all input variables and in the output variable: https://cran.r-project.org/...

View answer
Intuitive explanation of Fisher Information and Cramer-Rao bound
2 votes

This is the most intuitive article that I have seen so far: The Cramér-Rao Lower Bound on Variance: Adam and Eve’s “Uncertainty Principle” by Michael R. Powers, Journal of Risk Finance, Vol. 7, No. ...

View answer
Justification of one-tailed hypothesis testing
2 votes

The $p$-value is the probability of the respective event under the condition that $H_0$ is true. The simplest possible toy example are two coin tosses. The 2-sided $H_0$ would be that you consider the ...

View answer
Simple example that shows the advantages of Bayesian Model Averaging (BMA)
2 votes

A great resource for this is: Bayesian Model Averaging with BMS by Stefan Zeugner (2012) It is using the R-package BMS, more info can be found here: http://bms.zeugner.eu/ Two hands-on tutorials for ...

View answer
How to construct a rolling annual returns from a time series using R?
2 votes

You can use the zoo package: http://cran.r-project.org/web/packages/zoo/index.html You are very, very flexible to do all kinds of things, including aggregation and rolling functions: http://cran.r-...

View answer
Making sense of principal component analysis, eigenvectors & eigenvalues
1 votes

PCA basically is a projection of a higher-dimensional space into a lower dimensional space while preserving as much information as possible. I wrote a blog post where I explain PCA via the ...

View answer
Calculating payout of strategy for two one-armed bandits
Accepted answer
1 votes

First you construct the matrix with the transition probabilities, then you calculate the long run staying proportions via the normalized eigenvector of the biggest eigenvalue and last you weight this ...

View answer
Recurrent neural networks in R
1 votes

There is a new package out: rnn (on CRAN, on github), which implements a recurrent neural network in native R code. A nice example can be found here: http://firsttimeprogrammer.blogspot.de/2016/08/...

View answer
How to build scoring model (scorecard) from logistic regression?
Accepted answer
1 votes

The basic ideas are not that difficult: First model: You just multiply the respective coefficients with the new data points and see whether the sum is bigger than the negative intercept (then am is 1)...

View answer
Quantile transformation in R
1 votes

You can use the content method in the bin function in the OneR package for that. It works on vectors and dataframes. library(OneR) set.seed(2) df <- data.frame(a = rnorm(900), b = rnorm(900)) ...

View answer
Implementation of Convolutional Neural Network in R language
Accepted answer
1 votes

You can use the MXNetR package for that. The command is mx.symbol.Convolution() For the package documentation see: http://mxnet-bing.readthedocs.io/en/latest/R-package/index.html For a complete ...

View answer
Quadratic programming when the matrix is not positive definite
1 votes

You can build a workaround by using nearPD from the Matrix package like so: nearPD(D)$mat. nearPD computes the nearest positive definite matrix.

View answer
R: Finding relationships between 2 variables to determine any patterns in data
1 votes

Although you gave some data it is still hard to tell what would be the best method to use for your challenge. Still I give it a shot. In general it is always a good idea to do the steps you did ...

View answer
Does higher value of correlation between two values indicate it is good predictor?
1 votes

Concerning your question: "Does higher value of correlation between two values indicate it is good predictor?" In general I would be very cautious because one of the most important facts in ...

View answer
HMMFit underlying algorithms
1 votes

The RHmm-package has been deprecated. We use the depmixS4-package for our research and it is way better (although the learning curve is a little bit steeper due to the many options it has). You find ...

View answer
Predictive Modelling in R
Accepted answer
1 votes

This is a very broad and very basic question so I will recommend a very broad and basic book - but one that addresses your question thoroughly: Predictive Analytics For Dummies by Anasse Bari, ...

View answer
Term frequency/inverse document frequency (TF/IDF): weighting
1 votes

There is a new R package which can do this: textir: Inverse Regression for Text Analysis The relevant command is tfidf, the example from the manual: data(we8there) ## 20 high-variance tf-idf terms ...

View answer
A good resource to learn about the intuition and optimization models behind Markovitz modern portfolio theorem
1 votes

I find the books of Kritzman very intuitive - yet mathematically accurate: Kritzman (2000): Very gentle introduction Kritzman (2003): Goes a little deeper but still very understandable

View answer