New answers tagged

0

Your previous model won't be useful. In my opinion, you can not use it to make predictions in Covid Pandemic, since all bookings have had a huge decrease. I would recommend build a new model with only data of the Covid Pandemic time. Then, you will have a more accurate model for these times. The only problem if you have less historical data, but for sure, is ...


0

I canĀ“t give you an perfect answer because there is no code, dataset and the target what you want to achieve. Because the feature importances from random forest, is calculated based on the training data given to the model, not on predictions on a test dataset. That means, that is not the true prediction power. You should check, if there are difference on ...


1

If it's "find the house" then that implies that you need to find the house with those specific attributes in the dataset. I'm not sure what programming language you're using but it should be pretty simple to do such a thing anyways. If it's "find the price of a house with..." that implies there probably isn't such a house in the dataset ...


0

MLP typically refers to a type of neural network called a 'Multi-Layer Perceptron'. As you can read on the Wikipedia page, these neural networks consist of neurons organized in layers. Each neuron has a (typically fixed) function that transforms its weighted input to produce the output for that particular neuron known as its activation function. A 'network ...


0

I think it is a far question, since we should always try to translate any metric result to a useful interpretation. When talking about regression, we can judge the success of our model by checking whether that MAE or RMSE exceed what we are willing to accept. For instance, if we want to predict the temperature for the next days, and our result (in this case ...


2

According to the sklearn.svm.SVR documentation, the negative $R^2$ value indicates that your model is arbitrarily worse than the trend line on trainY. By default you should check the following: Does your model have a bias/intercept? If not you may observe negative $R^2$. Is testY derived from your training data? Am I using a linear function to fit the data? ...


0

Apart from the considerations about the quality of the data or whether or not the model is suitable for the problem, one good apporach is to try different combnations of the algorithm parameteres (using cross-validation) to come up with the best possible model. I mean, you can do a grid search or a randomizded search to find out which combination of the ...


2

A $R^2$ that is that low tells you that your model is not good. Therefore, you can both make it positive and nearer to 1 by : a) getting better/more data, or b) picking a better model for your data. Also, it'd be more helpful to plot the true/pred values against the underlying $X$ values and not just as a sequence.


1

I will use the "date" and "time" columns to pre-process your data and to construct your neural net input. RNN does not work well for very long-term dependancies... so, for example, creating a time series with all minutes in a month, won't probably work. You must select: How many samples your input data will have What is your sampling ...


0

I think your Batch size is too big - 1447. Try a small one like 4 and then increase it until you get OOM error again. Batch size always depends on the amount of your free GPU memory.


0

I want to know where Regression analysis is most used at Regression analysis is used for analyzing the relationship between some independent variables and a dependent variable. In particular it can be used for predicting/forecasting the dependent variable. In Machine Learning a regression task is a supervised task where the target (dependent variable) is ...


0

Any machine learning algorithm can be used as a weak learner for gradient boosting. Certain algorithms will work better and certain algorithms will work worse. Linear regression tries to minimize the sum of square errors (SSE). The first linear regression model might not leave any errors for subsequent linear regressions models to fit, thus there would be no ...


0

There are several options: Change the training dataset Use some of the test data as training data. This the best option since it better models the problem you are trying to solve. Since it happens over time, take only the most recent data for training. Manually engineer features. If you have knowledge of how the test data feature values are different, ...


1

As described in the error message, the problem is that your model needs more memory than your GPU has. Note that OOM stands for "out of memory". The specific layer that is demanding too much memory is Dense(154457, activation='relu'). Nevertheless, the last layer is even bigger. You should think if you really need an output of dimensionality 154457....


2

Essentially you want to pick a function that will give you the "size" of a matrix. The most obvious way I can think of is by choosing a matrix norm, which is a map $\lVert \cdot \rVert \colon \mathbb{R}^{k, k} \to [0, \infty)$ (or you could generalise to a complex $k \times k$ matrix if you wished). Your suggestion seems similar to computing $$S = \...


0

Cross-validation (CV) is a method meant essentially to accurately evaluate a model using some training data. As a consequence CV doesn't have to be used when training the final model, usually one simply uses the whole training data for that. Note that using one of the CV models would have two disadvantages: there's no reason to select one or the other (the ...


0

One option is scikit-learn's KBinsDiscretizer which bins continuous data into intervals.


0

Consider playing around with LASSO or Ridge-regressions, as these punish features with low predictive power. These are simple and strong methods for linear purposes. Your idea of using the feature importance from Random Forest could also be a suitable solution in cases of non-linearity.


0

Be careful choosing features based on correlation! Yes it is true that features that are correlated with the response variable may be good predictors, however if the features are correlated with each other then you are introducing multicollinearity into your model, which is bad. If you want to avoid this you should choose features which are correlated with ...


0

The pbounds must all be pairs; you cannot specify a list of options for max_depth. The package cannot deal with discrete hyperparameters very directly; see section 2, "Dealing with discrete parameters", of their "advanced tour" notebook about this.


1

In answer to your first question: The reason that your RMSE proceeded to increase as you increased the strength of your regularization (the value of $\lambda$) can be explained by reviewing the intuition behind what is happening when you increase the regularization of your model. Why did could my RMSE have kept increasing as I increased my regularization ...


1

The issue of using any linear model (a polynomial regression is a particular case of a linear model, with polynomial features), is that an ensemble of linear models is still a linear model. So, the family of models to optimize from given by the boosted polynomial regression and the single polynomial regression are the same. This doesn't happen with trees, as ...


1

The main problem in your test is that your X has the same scale as your noise level (0-1), as a result, adding a noise changes your data distribution significantly. This is your data distribution before and after adding noise. It is like the noise is 50-200% more than your initial data. That's why you get a better result with CV than the "ground truth ...


0

I think your question depends almost entirely on the actual problem that you're trying to represent, i.e. it depends on (your) expert knowledge. Yes, data generated this way should be considered synthetic. However there's no general requirement to add noise: adding noise (as well as how much and in which way) is normally meant to make synthetic data more ...


0

I am running into exactly this problem. I am looking at correlations between load and throughput. Literally (load,throughput) pairs. But if you are measuring a real system that is usually at load 1000, you may never get data for load in low values like load 1,2,3, etc. (ie: what is Google throughput at 1 user only? We will never get data for it.) So, ...


0

You can look at this article. Helmreich, James & Pruzek, Robert. (2008). PSAgraphics: An R Package to Support Propensity Score Analysis. Journal of Statistical Software. 29. 10.18637/jss.v029.i06. In section three, they show an example of how you can estimate propensity scores and stratify the data based on them. There is also a good example in Valliant, ...


0

If the model is not derived from the data then it must be built manually, so non data driven means rule based. This was the big trend in AI in the 80s before Machine Learning, these pre-ML automatic prediction systems were called expert systems and they were quite successful at the time in industry (here are some examples of applications). The way one would ...


Top 50 recent answers are included