I've been following a lot of tutorials using lstms to forecast timeseries data. My question is that how do we predict on new data that is not part of the dataset since almost all the tutorials show the predict function in Keras being used on the test data split.
How do we actually forecast into the future?
Usually, you create your training data such that the model receives n points and predict the following m points. Once you have your model trained, you take the last n available points of your dataset or new points from the present, and the model will output a prediction of m points in the future.
If you want to predict more than m points in the future, you could predict m points and use it as input to predict another m points, and so on. However, you should be aware that using this technique you will probably get worse results as you are accumulating errors.
Related
I'm fairly new to data analysis and machine learning. I've been carrying out some KNN classification analysis on a breast cancer dataset in python's sklearn module. I have the following code which attemps to find the optimal k for classification of a target variable.
from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
import matplotlib.pyplot as plt
breast_cancer_data = load_breast_cancer()
training_data, validation_data, training_labels, validation_labels = train_test_split(breast_cancer_data.data, breast_cancer_data.target, test_size = 0.2, random_state = 40)
results = []
for k in range(1,101):
classifier = KNeighborsClassifier(n_neighbors = k)
classifier.fit(training_data, training_labels)
results.append(classifier.score(validation_data, validation_labels))
k_list = range(1,101)
plt.plot(k_list, results)
plt.ylim(0.85,0.99)
plt.xlabel("k")
plt.ylabel("Accuracy")
plt.title("Breast Cancer Classifier Accuracy")
plt.show()
The code loops through 1 to 100 and generates 100 KNN models with 'k' set to incremental values in the range 1 to 100. The performance of each of those models is saved to a list and a plot is generated showing 'k' on the x-axis and model performance on the y-axis.
The problem I have is that when I change the random_state parameter when spliting the data into training and testing partitions this results in completely different plots indicating varying model performance for different 'k'values for different dataset partitions.
For me this makes it difficult to decide which 'k' is optimal as the algorithm performs differently for different 'k's using different random states. Surely this doesn't mean that, for this particular dataset, 'k' is arbitrary? Can anyone help shed some light on this?
Thanks in anticipation
This is completely expected. When you do the train-test-split, you are effectively sampling from your original population. This means that when you fit a model, any statistic (such as a model parameter estimate, or a model score) will it self be a sample estimate taken from some distribution. What you really want is a confidence interval around this score and the easiest way to get that is to repeat the sampling and remeasure the score.
But you have to be very careful how you do this. Here are some robust options:
1. Cross Validation
The most common solution to this problem is to use k-fold cross-validation. In order not to confuse this k with the k from knn I'm going to use a capital for cross-validation (but bear in mind this is not normal nomenclature) This is a scheme to do the suggestion above but without a target leak. Instead of creating many splits at random, you split the data into K parts (called folds). You then train K models each time on K-1 folds of the data leaving aside a different fold as your test set each time. Now each model is independent and without a target leak. It turns out that the mean of whatever success score you use from these K models on their K separate test sets is a good estimate for the performance of training a model with those hyperparameters on the whole set. So now you should get a more stable score for each of your different values of k (small k for knn) and you can choose a final k this way.
Some extra notes:
Accuracy is a bad measure for classification performance. Look at scores like precision vs recall or AUROC or f1.
Don't try program CV yourself, use sklearns GridSearchCV
If you are doing any preprocessing on your data that calculates some sort of state using the data, that needs to be done on only the training data in each fold. For example if you are scaling your data you can't include the test data when you do the scaling. You need to fit (and transform) the scaler on the training data and then use that same scaler to transform on your test data (don't fit again). To get this to work in CV you need to use sklearn Pipelines. This is very important, make sure you understand it.
You might get more stability if you stratify your train-test-split based on the output class. See the stratify argument on train_test_split.
Note the CV is the industry standard and that's what you should do, but there are other options:
2. Bootstrapping
You can read about this in detail in introduction to statistical learning section 5.2 (pg 187) with examples in section 5.3.4.
The idea is to take you training set and draw a random sample from it with replacement. This means you end up with some repeated records. You take this new training set, train and model and then score it on the records that didn't make it into the bootstrapped sample (often called out-of-bag samples). You repeat this process multiple times. You can now get a distribution of your score (e.g. accuracy) which you can use to choose your hyper-parameter rather than just the point estimate you were using before.
3. Making sure you test set is representative of your validation set
Jeremy Howard has a very interesting suggestion on how to calibrate your validation set to be a good representation of your test set. You only need to watch about 5 minutes from where that link starts. The idea is to split into three sets (which you should be doing anyway to choose a hyper parameter like k), train a bunch of very different but simple quick models on your train set and then score them on both your validation and test set. It is OK to use the test set here because these aren't real models that will influence your final model. Then plot the validation scores vs the test scores. They should fall roughly on a straight line (the y=x line). If they do, this means the validation set and test set are both either good or bad, i.e. performance in the validation set is representative of performance in the test set. If they don't fall on this straight line, it means the model scores you get from you validation set are not indicative of the score you'll get on unseen data and thus you can't use that split to train a sensible model.
4. Get a larger data set
This is obviously not very practical for your situation but I thought I'd mention it for completeness. As your sample size increases, your standard error drops (i.e. you can get tighter bounds on your confidence intervals). But you'll need more training and more test data. While you might not have access to that here, it's worth keeping in mind for real world situations where you can assess the trade-off of the cost of gathering new data vs the desired accuracy in assessing your model performance (and probably the performance itself too).
This "behavior" is to be expected. Of course you get different results, when training and test is split differently.
You can approach the problem statistically, by repeating each 'k' several times with new train-validation-splits. Then take the median performance for each k. Or even better: look at the performance distribution and the median. A narrow performance distribution for a given 'k' is also a good sign that the 'k' is chosen well.
Afterwards you can use the test set to test your model
I know this may be a basic question but I want to know if I am using the train, test split correctly.
Say I have data that ends at 2019, and I want to predict values in the next 5 years.
The graph I produced is provided below:
My training data starts from 1996-2014 and my test data starts from 2014-2019. The test data perfectly fits the training data. I then used this test data to make predictions from 2019-2024.
Is this the correct way to do it, or my predictions should also be from 2014-2019 just like the test data?
The test/validation data is useful for you to evaluate the predictor to use. Once you have decided which model to use, you should train the model with the whole dataset 1996-2019 so that you do not lose possible valuable knowledge from 2014-2019. Take into account that when working with time-series, usually the newer part of the serie has more importance in your prediction than older values of the serie.
I am trying to build a model that predicts the shipping volume of each month, week, and day.
I found that the decision tree-based model works better than linear regression.
But I read some articles about machine learning and it says decision tree based model can't predict future which model didn't learn. (extrapolation issues)
So I think it means that if the data is spread between the dates that train data has, the model can predcit well, but if the date of data is out of the range, it can not.
I'd like to confirm if my understand is correct.
some posting shows prediction for datetime based data using random forest model, and it makes me confused.
Also please let me know if there is any way to overcome extrapolation issues on decision tree based model.
It depends on the data.
Decision tree predicts class value of any sample in range of [minimum of class value of training data, maximum of class value of training data]. For example, let there are five samples [(X1, Y1), (X2, Y2), ..., (X5, Y5)], and well trained tree has two decision node. The first node N1 includes (X1, Y1), (X2, Y2) and the other node N2 includes (X3, Y3), (X4, Y4), and (X5, Y5). Then the tree will predict a new sample as mean of Y1 and Y2 when the sample reaches N1, but it will predict a new sample as men of Y3, Y4, Y5 when the sample reaches N2.
With this reason, if the class value of new sample could be bigger than the maximum of class value of training data or could be smaller than the minimum of class value of training data, it is not recommend to use decision tree. Otherwise, tree-based model such as random forest shows good performance.
There can be different forms of extrapolation issues here.
As already mentioned a classical decision tree for classification can only predict values it has encountered in its training/creation process. In that sense you won't predict any previously unseen values.
This issue can be remedied if you have the classifier predict relative updates instead of absolute values. But you need to have some understanding of your data, to determine what works best for different cases.
Things are similar for a decision tree used for regression.
The next issue with "extrapolation" is that decision trees might perform badly if your training data has changing statistics over time. Again, I would propose to predict update relationships.
Otherwise, predictions based on training data from a more recent past might yield better predictions. Since individual decision trees can't be trained in an online manner, you would have to create a new decision tree every x time steps.
Going further than this I'd say you'll want to start thinking in state machines and trying to use your classifier for state predictions. But this a fairly uncharted domain of theory for decision trees from when I last checked. This will work better if you already have some for of model for your data relationships in mind.
I am using z-score to normalize my data before training my model. When I do predictions on a daily basis, I tend to have very few observations each day, perhaps just a dozen or so. My question is, can I normalize the test data just by itself, or should I attach it to the entire training set to normalize it?
The reason I am asking is, the normalization is based on mean and std_dev, which obviously might look very different if my dataset consists only of a few observations.
You need to have all of your data in the same units. Among other things, this means that you need to use the same normalization transformation for all of your input. You don't need to include the new data in the training per se -- however, keep the parameters of the normalization (the m and b of y = mx + b) and apply those to the test data as you receive them.
It's certainly not a good idea to predict on a test set using a model trained with a very different data distribution. I would use the same mean and std of your training data to normalize you test set.
Let's say that I have a data file like:
Index,product_buying_date,col1,col2
0,2013-01-16,34,Jack
1,2013-01-12,43,Molly
2,2013-01-21,21,Adam
3,2014-01-09,54,Peirce
4,2014-01-17,38,Goldberg
5,2015-01-05,72,Chandler
..
..
2000000,2015-01-27,32,Mike
with some more data and I have a target variable y. Assume something as per your convenience.
Now I am aware that we divide the data into 2 parts i.e. Train and Test. And then we divide Train into 70:30, build the model with 70% and validate it with 30%. We tune the parameters so that model does not get overfit. And then predict with the Test data. For example: I divide 2000000 into two equal parts. 1000000 is train and I divide it in validate i.e. 30% of 1000000 which is 300000 and 70% is where I build the model i.e. 700000.
QUESTION: Is the above logic depending upon how the original data splits?
Generally we shuffle the data and then break it into train, validate and test. (train + validate = Train). (Please don't confuse here)
But what if the split is alternate. Like When I divide it in Train and Test first, I give even rows to Test and odd rows to Train. (Here data is initially sort on the basis of 'product_buying_date' column so when i split it in odd and even rows it gets uniformly split.
And when I build the model with Train I overfit it so that I get maximum AUC with Test data.
QUESTION: Isn't overfitting helping in this case?
QUESTION: Is the above logic depending upon how the original data
splits?
If dataset is large(hundred of thousand), you can randomly split the data and you should not have any problem but if dataset is small then you can adopt the different approaches like cross-validation to generate the data set. Cross-validation states that you split you make n number of training-validation set out of your Training set.
suppose you have 2000 data points, you split like
1000 - Training dataset
1000 - testing dataset.
5-cross validation would mean that you would make five 800/200 training/validation dataset.
QUESTION: Isn't overfitting helping in this case?
Number one rule of the machine learning is that, you don't touch the test data set. It's a holly data set that should not be touched.
If you overfit the test data to get maximum AUC score then there won't be any meaning of validation dataset. Foremost aim of any ml algorithm is to reduce the generalization error i.e. algorithm should be able to perform good on unseen data. If you would tune your algorithm with testing data. you won't be able to meet this criteria. In cross-validation also you do not touch your testing set. you select your algorithm. tune its parameter with validation dataset and after you have done with that apply your algorithm to test dataset which is your final score.