ensemble results from different classifier in R [closed] - r

Closed. This question needs details or clarity. It is not currently accepting answers.
Want to improve this question? Add details and clarify the problem by editing this post.
Closed 7 years ago.
Improve this question
I have prediction of my data in different classifier. I would like to ensemble the results of them in order to get better final result. Is it possible in R?
lets say:
SVMpredict=[1 0 0 1...]
RFpredict=[1 1 0 1 ...]
NNpredict=[0 0 0 1 ...]
is it possible to combine results by any ensemble technique in R?how?
thanks
Edited:
I run my classifiers on different samples (my case DNA chromosomes). In some samples SVM works better than the others like RF. I want a technique to ensemble the results by considering which classifier works better.
for example if i take average of output probabilities and round them, it would be considered as all classifier are equal effective in result. but when SVM worked better, we should consider results for SVM (with 86% accuracy) has 60% importance and 25% (72% accuracy) for RF and 15% NN (64% accuracy). (these numbers are just examples for clarification)
is there anyway that I can do that?

It depends on the structure of the output of Your classifier.
If it is {0,1} outcome, as You provided, You can just make mean of the predictions and then average it and round it:
round((SVMpredict+RFpredict+NNpredict)/3)
If You know performance of the classifiers, weighted mean is a good idea - favor the ones that perform better. Hardcore apporach is to optimize weights via optim function.
If You know class probabilites for each prediction, it is better average them instead of letting them just vote ({0,1} output case above).

Related

How should I select features for logistic regression in R? [closed]

Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about programming within the scope defined in the help center.
Closed 1 year ago.
Improve this question
I tried several ways of selecting predictors for a logistic regression in R. I used lasso logistic regression to get rid of irrelevant features, cutting their number from 60 to 24, then I used those 24 variables in my stepAIC logistic regression, after which I further cut 1 variable with p-value of approximately 0.1. What other feature selection methods I can or even should use? I tried to look for Anova correlation coefficient analysis, but I didn't find any examples for R. And I think I cannot use correlation heatmap in this situation since my output is categorical? I seen some instances recommending Lasso and StepAIC, and other instances criticising them, but I didn't find any definitive comprehensive alternative, which left me confused.
Given the methodological nature of your question you might also get a more detailed answer at Cross Validated: https://stats.stackexchange.com/
From your information provided, 23-24 independent variables seems quite a number to me. If you do not have a large sample, remember that overfitting might be an issue (i.e. low cases to variables ratio). Indications of overfitting are large parameter estimates & standard errors, or failure of convergence, for instance. You obviously have already used stepwise variable selection according to stepAIC which would have also been my first try if I would have chosen to let the model do the variable selection.
If you spot any issues with standard errors/parameter estimates further options down the road might be to collapse categories of independent variables, or check whether there is any evidence of multicollinearity which could also result in deleting highly-correlated variables and narrow down the number of remaining features.
Apart from a strictly mathematical approach you might also want to identify features that are likely to be related to your outcome of interest according to your underlying content hypothesis and your previous experience, meaning to look at the model from your point of view as expert in your field of interest.
If sample size is not an issue and the point is reduction of feature numbers, you may consider running a principal component analysis (PCA) to find out about highly correlated features and do the regression with the PCAs instead which are non-correlated linear combination of your "old" features. A package to accomplish PCA is factoextra using prcomp or princomp arguments http://www.sthda.com/english/articles/31-principal-component-methods-in-r-practical-guide/118-principal-component-analysis-in-r-prcomp-vs-princomp/

DESeq2 design matrix including RIN as covariate in the formula [closed]

Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about programming within the scope defined in the help center.
Closed 2 years ago.
Improve this question
I have been following the last DESeq2 pipeline to perform an RNAseq analysis. My problem is the rin of the experimental samples is quite low compared to the control ones. Iread a paper in which they perform RNAseq analysis with time-course RNA degradation and conclude that including RIN value as a covariate can mitigate some of the effects of low rin in samples.
My question is how I should construct the design in the DESeq2 object:
~conditions+rin
~conditions*rin
~conditions:rin
none of them... :)
I cannot find proper resources where explain how to construct these models (I am new to the field...) and I recognise I crashed against a wall with these kinds of things. I would appreciate also some links to good resources to be able to understand which one is correct and why.
Thank you so much
Turns out to be quite long for typing in a comment.
It depends on your data.
First of all, counts ~conditions:rin does not make sense in your case, because conditions is categorical. You cannot fit only an interaction term model.
I would go with counts ~condition + rin, this assumes there is a condition effect and a linear effect from rin. And the counts' dependency of rin is independent of condition.
As you mentioned, rin in one of the conditions is quite low, but is there any reason to suspect the relationship between rin and counts to differ in the two conditions? If you fit counts ~condition * rin, you are assuming a condition effect and a rin effect that is different in conditions. Meaning a different slope for rin effect if you plot counts vs rin. You need to take a few genes out, and see whether this is true. And also, for fitting this model, you need quite a lot of samples to estimate the effects accurately. See if both of these holds

Machine Learning Classification with only binary numbers [closed]

Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 5 years ago.
Improve this question
I have 50 predictors and 1 target variable. All my predictors and target variable are only binary numbers 0s and 1s. I am performing my analysis using R.
I will be implementing four algorithms.
1. RF
2. Log Reg
3. SVM
4. LDA
I have the following questions:
I convert them all into factors. How should i treat my variables priorly, before feeding them into my other algorithms.
I used the caret package to train my model, it takes very much time. I do practice ML regularly, but I dont know how to proceed with all variable being binary.
How to remove collinear variables?
I'm not mostly R-user, but Python. Bet there is common approach:
1. Check you columns. Remove column if number of zeroes or ones is > 95% of total amount (you can try 2.5% or even 1% later).
2. Run simple Random Forest by default and get feature importance. Columns that are unnecessary you can process with LDA.
3. Check target column. If it's highly unbalanced try oversampling or downsampling. Or use classification methods that can handle unbalanced target column (like XGBoost).
For Linear regression you'll need to calculate correlation matrix and remove correlated columns. Other methods can live without it.
Please check SVM (or SVC) does it support all features to be boolean or not. But usually it works very good with binary classification.
Also I advice to try Neural Network.
PS About collinear variables. I wrote a code on Python for my project. That's simple - you can do it:
- plot correlation matrix
- find pairs that have correlation over some threshold
- remove column that have lower correlation with target variable (you can also check that columns you want to remove is not important, otherwise try other way, probably union columns)
In my code I ran this algorithm iteratively for different thresholds: from 0.99 down to 0.9. Works good.

Test for significance in a time series using R [closed]

Closed. This question is off-topic. It is not currently accepting answers.
Want to improve this question? Update the question so it's on-topic for Stack Overflow.
Closed 10 years ago.
Improve this question
Given a simplified example time series looking at a population by year
Year<-c(2001,2002,2003,2004,2005,2006)
Pop<-c(1,4,7,9,20,21)
DF<-data.frame(Year,Pop)
What is the best method to test for significance in terms of change between years/ which years are significantly different from each other?
As #joran mentioned, this is really a statistics question rather than a programming question. You could try asking on http://stats.stackexchange.com to obtain more statistical expertise.
In brief, however, two approaches come to mind immediately:
If you fit a regression line to the population vs. year and have a statistically significant slope, that would indicate that there is an overall trend in population over the years, i.e. use lm() in R, like this lmPop <- lm(Pop ~ Year,data=DF).
You could divide the time period into blocks (e.g. the first three years and the last three years), and assume that the population figures for the years in each block are all estimates of the mean population during that block of years. That would give you a mean and a standard deviation of the population for each block of years, which would let you do a t-test, like this: t.test(Pop[1:3],Pop[4:6]).
Both of these approaches suffer from some potential difficulties and the validity of each would depend on the nature of the data that you're examining. For the sample data, however, the first approach suggests that there appears to be a trend over time at a 95% confidence level (p=0.00214 for the slope coefficient) while the second approach suggests that the null hypothesis that there is no difference in means cannot be falsified at the 95% confidence level (p = 0.06332).
They're all significantly different from each other. 1 is significantly different from 4, 4 is significantly different from 7 and so on.
Wait, that's not what you meant? Well, that's all the information you've given us. As a statistician, I can't work with anything more.
So now you tell us something else. "Are any of the values significantly different from a straight line where the variation in the Pop values are independent Normally distributed values with mean 0 and the same variance?" or something.
Simply put, just a bunch of numbers can not be the subject of a statistical analysis. Working with a statistician you need to agree on a model for the data, and then the statistical methods can answer questions about significance and uncertainty.
I think that's often the thing non-statisticians don't get. They go "here's my numbers, is this significant?" - which usually means typing them into SPSS and getting a p-value out.
[have flagged this Q for transfer to stats.stackexchange.com where it belongs]

Power Analysis in [R] for Two-Way Anova [closed]

Closed. This question is off-topic. It is not currently accepting answers.
Want to improve this question? Update the question so it's on-topic for Stack Overflow.
Closed 12 years ago.
Improve this question
I am trying to calculate the necessary sample size for a 2x2 factorial design. I have two questions.
1) I am using the package pwr and the one way anova function to calculate the necessary sample size using the following code
pwr.anova.test(k = , n = , f = , sig.level = , power = )
However, I would like to look at two way anova, since this is more efficient at estimating group means than one way anova. There is no two-way anova function that I could find. Is there a package or routine in [R] to do this?
2) Moreover, am I safe in assuming that since I am using a one-way anova power calculations, that the sample size will be more conservative (i.e. larger)?
In a 2 x 2 ANOVA involving Factor A, Factor B, and AxB, you will get separate statistical power estimates for each of these three effects.
G Power 3 provides free software and some clear tutorials for estimating power of effects in factorial designs:
http://www.psycho.uni-duesseldorf.de/abteilungen/aap/gpower3/user-guide-by-design
After searching - I couldn't find any solution for this online.
What I would suggest you to do (if you know how) is to program this using a simulation.
If you don't know how to do it, then write a SO question about "How can I write a simulation of two-way anova, to achieve power analysis" and see what people might help you with :)
Also, you could start by reviewing the code here:
http://www.rforge.net/doc/packages/NCStats/power.sim.html
For a start on power calculation through simulation.
Notice what Jeromy wrote - this power analysis is for multiple outcomes.
Interesting subject - I'd love to followup on it.
Best,
Tal

Resources