Reasons to not use PCA for feature selection

https://blog.kxy.ai/5-reasons-you-should-never-use-pca-for-feature-selection/

by leonry • 4 years ago

99 85 4 years ago

blog.kxy.ai

jstx16• 4 years ago

The only reason you need: PCA is not a feature selection algorithm

Is the author misunderstanding something very basic or are they deliberately writing this way for clicks and attention? I can see that they have great credentials so probably the latter? It's a weird article.

bllguo4• 4 years ago

practically speaking i have never had a reason to do this with PCA, but i find this reaction very weird. i have encountered the concept of using PCA to reduce the dimensionality of your data before training many times, including in university. cursory searches show plenty of discussion, questions on crossvalidated, etc.

jstx16• 4 years ago

My point is that dimensionality reduction != feature selection. PCA is dimensionality reduction, not feature selection.

j16sdiz2• 4 years ago

PCA as feature selection is quite popular. Many text book on classification teach this.

rlayton22• 4 years ago

So you have an example? I've never seen this and would like to see what people are doing here. I teach a lot of newbies data mining, so I'm very interested in how people get it wrong

rlayton22• 4 years ago

Yeah I thought I was reading the title wrong!

For anyone that isn't aware, the role of pca is to create new (synthetic) features that represent the original features.

It does not tell you which features of the original set are good for feature selection purposes.

pas2• 4 years ago

so what's a good way/algorithm/strategy/method/technique to select features? (obviously asking for a friend who's not that familiar with this, whereas I've all the black belts in feature engineering!)

platz5• 4 years ago

Isn't the article in fact saying that the new synthetic features are not good features for training?

galangalalgol• 4 years ago

The eigen values do give information as to energy, so if your selection criteria is simply to pick the linear component with the most energy, you can use PCA to select that feature and extract it with the corresponding eigen vector. The MUSIC algorithm is a classic example.

Edit: having now actually read the article, the case I mention falls into the author's test of (do linear combinations of my features make sense and have as much of a relationship to the target as the features themselves).

pas2• 4 years ago

Could you elaborate on this a bit? Are there feature selection algorithms?

jstx16• 4 years ago

PCA is a form for dimensionality reduction. It can be used to reduce your data to a only few features while preserving some properties like the predictive quality of your model. Simple example: you give PCA 10 input features, it gives back the principal components, you pick the first 2 principal components and use them in your regression model (you still select how many componenets to use yourself, PCA doesn't do that for you).

Feature selection is well... selecting features. You start with your original 10 features and you pick which ones to use. You can do this based on domain knowledge, based on analysing variance, based on correlation between features, based on feature importance in some model, by adding/removing/shuffling features and seeing how your model performs etc.

weberer• 4 years ago

Are there feature selection algorithms?

Yeah. Here's a few

https://scribe.citizen4.eu/feature-selection-how-to-throw-aw...

derbOac• 4 years ago

I think you could think of PCA as a feature selection candidate when the prediction target is unknown or random?

it_does_follow2• 4 years ago

It's very clear if you read the article that what the author is calling "feature selection" might be better termed "feature generation". He explicitly calls out what he means in the post:

When used for feature selection, data scientists typically regard z^p:=(z_1,…,z_p) as a feature vector than contains fewer and richer representations than the original input x for predicting a target y.

I don't even think this is necessarily incorrect terminology, especially given the author's background of working primarily for Google and the like. It's the difference between considering feature section as "choosing from a list of the provided features" vs "choosing from the set of all possible features". The author's term makes perfect sense given the latter.

PCA is used for this all the time in the field. There have been an astounding number of presentations I've seen where people start with PCA/SVD as the first round of feature transformation. I always ask "why are you doing that?" and the answer is always mumbling with shoulder shrugging.

This is a solid post and I find it odd that you try to dismiss it as either ignorant or click bait, when a quick skim of it dismisses both of these options.

anonymoushn• 4 years ago

Much of the text is outside the viewport and cannot be scrolled to on mobile.

raverbashing• 4 years ago

Here's a better idea, don't use generalized statements but rather test your data first.

If your models fits well (and doesn't overfit) after PCA, then go for it. If not, revisit.

PCA has its place, and as the other commenter said, sure, it's not a feature selection algorithm. Or you can just feature select manually.

shreyshnaccount4• 4 years ago

You might not even need feature selection, just a dimensionality reduction in some cases

platz5• 4 years ago

1) Does linearly combining my features make any sense?
2) Can I think of an explanation for why the linearly combined features could have as simple a relationship to the target as the original features?

the article provided a negative example where pca does not fit, but doesnt provide an example where it does or what pca is actually used for. I come away from this article thinking pca is useless.

what would be an example where 2) is true?

I cannot answer 2) without already having experience of what explanations there could possibly be. ( 2) is almost begging the question, at least pedagogically- pca is good when the features are good for pca)

When does linearly combining my features "make sense"? again, an example is not provided

ThereIsNoWorry• 4 years ago

I mean, that's the point. What kind of real world complexities are actually linearly explainable? Almost none. It goes all the way back to why classical statistics failed to provide real world value after a certain point and the trend has been going towards using non-linear black boxes for the last decade.

platz5• 4 years ago

So you basically are saying PCA is a useless procedure.

Also, "after a certain point" is doing a lot of heavy lifting in that sentance.

hervature4• 4 years ago

The article is silly because PCA cannot select features. It is all about dimensionality reduction. You should think of PCA as the equivalent to VAEs from the neural network world. The idea would be something like this: you have big images (let's say 4k) and this is too expensive to train with/store forever. So, you collect a training set, train a PCA on these images, and then you can convert your 4k images to 720p or even 10 numbers, which you then use to predict/train whatever you want. Of course, we have algorithms that scale images but maybe all your images are of cats and there is a specific linear transformation that contains more information from the 4k image than simply scaling. The implicit thing here is that you still are collecting 4k images but just immediately compress them down using your trained PCA transformation.

So, although you have less numbers than before, you still need to collect the original data. A real feature selection process would be able to do something like: "the proximity of the closest Applebees is not important to predict house prices, you should probably stop wasting your time calculating this number". As others have mentioned, L1 regression or some statistical procedure to identify useless features is typically how this is done. I would also add that domain knowledge is probably your #1 feature selection because we have to restrict the variables we input in the first place and which data we prioritize is inherently selecting the features.

dkersten• 4 years ago

So would this be a more-or-less correct tl;dr:

Dimensionality reduction is a compressing data in a way that retains the most important information for the task

Feature selection is removing unimportant information (keeping/collecting, or selecting, only the important parts)

Both cut down on the amount of data you end up with, but one does it by finding a representation that is smaller, the other does it by discarding unnecessary data (or, rather, telling you which data is necessary, so you can stop collecting the unnecessary data).

mrow84• 4 years ago

One method is to drop variables with small coefficients in the top n principal components. This is feature selection, by your definition.

rybosworld• 4 years ago

PCA has a niche use-case. It's more often harmful than not in my experience.

shreyshnaccount4• 4 years ago

Can you elaborate? What other techniques do you find useful or better for either feature extraction or dimensionality reduction?

time_to_smile• 4 years ago

The entire field of deep learning is essentially performing non-linear dimensionality reduction, ultimately projecting the data on to a manifold where it is linearly separable.

The last layer of most neural networks is essentially just logistic regression, so if you take that final layer before the soft max you have a lower dimensional representation of the original data.

You can extend this things to more sophisticate example where you're not just learning a target but some other more general objective. For example you might want to learn a representation that minimizes the cosine distance between objects in a similar category.

PCA is only useful for compression (which is rarely a problem in contemporary applications) as far as pure feature engineering goes. However when you realize that PCA is the same as a linear auto-encoder then it does serve as a good intellectual starting point for more sophisticated dimensionality reduction techniques.

quanto• 4 years ago

PCA can be construed as a loss compression of the data matrix. In fact, Eckart-Young theorem shows that this is an optimal compression (optimal w.r.t low rank, i.e. space to hold the values). In the language of OP, this shows the minimal energy loss for a given space constraint.

The key word is "lossy". It may as well be that the loss'ed part had the signal for further classification down the pipeline. Or may be not. It depends on the case.

srean• 4 years ago

There was a recent discussion on PCA for classification that I had walked into, however, everyone had left the building when I joined. Since I run into this conceptual misunderstanding of PCA's relevance to classification often, let me repeat what I said there.

The problem with using PCA as a preprocessing step for linear classification is that this dimensionality reduction step is being done without paying any heed to the end goal -- better linear separation of the classes. One can get lucky and get a low-d projection that separates well but that is pure luck. Let me see if I can draw an example

           ++++++
       ++++++++++++++++
             +++
           ---------
       -----------------
            -------

The '+' and '-' denote the data points of the two different classes. In this example the PCA direction will be the along the X axis, which would be the worst axis to project on to separate the classes. The best in this case would have been the Y axis.

A far better approach would be to use a dimensionality reduction technique that is aware of the end goal. One such example is Fisher discriminant analysis and its kernelized variant.

shreyshnaccount4• 4 years ago

The answer, as always, is IT DEPENDS.

civilized4• 4 years ago

Now that we've all said "feature selection is not dimensionality reduction" to our hearts' content, could we return to the point of the article?

Regardless of whether you're doing feature selection or dimensionality reduction, the point remains that, if you're doing supervised learning, PCA is just compressing your X space, without any regard to your y. It could be that the last principal component of X, containing only 0.1% of the variance, contains 100% of the correlation between X and y.

Using PCA for dimensionality reduction in a supervised learning context means throwing out an unknown amount of signal, which could be up to 100% of the signal.

Now for unsupervised, exploratory analysis, PCA is definitely a candidate, but there are plenty of often-better alternatives there too.

jstx16• 4 years ago

For me PCA is strictly in a bucket labeled "this comes up a lot in data science interviews but you never actually need to use it at work".

civilized4• 4 years ago

For me the contents of this bucket are things like SVM, Naive Bayes, and k-means. PCA is actually in a slightly higher tier than those three.

it_does_follow2• 4 years ago

You've never had to any kind of factor analysis in your work or done any searching for latent variables that map to customer/stakeholder question? Given the number of people I've worked with that are interested in modeling "engagement", I find this hard to believe.

PCA is an incredibly valuable tool that I've used in most jobs I've had. It's just a terrible idea as a default part of a feature engineering pipeline (which is what the author is talking about in terms of "feature selection"), for reasons outline in this article.

I suggest you don't be quite so quick to dismiss important concepts in this area, and before criticizing this post, at least read through it (I noticed your comment about misunderstand what the author is discussing by "feature selection" is the top comment here).

hatmatrix• 4 years ago

If you're looking for the part of the signal in X that covaries with y, the first thing to try is partial least squares.

mturmon• 4 years ago

Yes. In the simplified linear setting we're considering...

If you want to do dimension reduction and only have X, use PCA.

If you want to do dimension reduction and also have a "y", consider PLS or CCA (canonical correlation analysis).

Some background: https://www.cs.cmu.edu/~tom/10701_sp11/slides/CCA_tutorial.p...

ylks6• 4 years ago

You nailed it.

first_post• 4 years ago

What would you recommend for feature selection in say, single-cell RNA seq studies (Typical dataset is ~10,000 x ~30000 (cells x genes) with >90% of your table filled with 0s (which could be due to biological or technical noise)

PCA and UMAP are yes, dimensionality reduction methods, but are often seen as tools for feature selection.

See slide 61 Here: https://physiology.med.cornell.edu/faculty/skrabanek/lab/ang...

plandis• 4 years ago

I followed the math but I don’t know ML… Do practitioners really use “energy” and “conservation of energy”? That just seems overly confusing.

dekhn2• 4 years ago

I've never heard the term used before in ML or statistical modelling.

ylks6• 4 years ago

Energy is a shorthand for L2 norm (in a probabilistic sense). This similar to the definition of the energy of a signal in signal processing. 'Information content' is an alternative for 'energy' here, but it can be mistaken for entropy.

ylks6• 4 years ago

Hello,

I'm the author of the post. I'm slightly late to the party, but I'll try to clarify a few misunderstandings.

First and foremost, the post deals with the following scenario too many data scientists find themselves in: "I have (generated) a lot of features; let me do PCA and train my model using the top few principal components". This is a terrible idea and the post explains why.

Second, there seems to be a debate about 'feature selection' vs. 'feature construction' (or 'feature generation'), and whether PCA is of the former or latter type. Here are the definitions I use in the whole blog.

Feature Construction is the process consisting of generating candidate features (i.e. transformations of the original inputs) that might have a simpler relationship with the target, one that models in our toolbox can reliably learn.

E.g. a linear model cannot learn a quadratic function. However, because a quadratic function of x is linear in [x, x^2], the feature transformation x -> [x, x^2] is needed to make our quadratic function learnable by a linear model.

Check out this post for more details on what's at stake during feature construction: https://blog.kxy.ai/feature-construction/.

Feature Selection is the process consisting of removing useless features in the set of candidates generated by feature construction. A feature is deemed useless when it is uninformative about the target or redundant.

Check out this post for a formulation of three key properties of a feature using Shapley values: feature importance, feature usefulness, and feature potential. https://blog.kxy.ai/feature-engineering-with-game-theory-bey...

In the scenario the blog post deals with (i.e. "I have (generated) a lot of features; let me do PCA and train my model using the top few principal components"), data scientists do both feature construction (full PCA, i.e. projecting the original input onto eigenvectors to obtain as many principal components as the dimension of the original input) AND feature selection (only selecting the first few principal components with the highest eigenvalues).

When the goal is to predict y from x, using PCA for either feature construction OR feature selection is a bad idea!

For feature construction, there is nothing in PCA that will intrinsically guarantee that a linear combination of coordinates of x will have a simpler relationship to y than x itself. PCA does not even use y in this case! E.g. Imagine all coordinates of x but the first are pure noise (as far as predicting y is concerned). Any linear combination of x will just make your inputs noisy!

For feature selection, even assuming principal components make sense as features, principal components with the highest variances (i.e. corresponding to the highest eigenvalues) need not be the most useful for predicting y! High variance does not imply high signal.

Reasons to not use PCA for feature selection

Related Stories