# kernel method linear regression

dezembro 21, 2020 3:38 am Deixe um comentário

, i Unit 5: Kernel Methods. Kernel method buys us the ability to handle nonlinearity. d x ) K Exercice 3: (check the solution) Implement the ISTA algorithm, display the convergence of the energy. \newcommand{\de}{\delta} j ( ^ ∑ x 28 Kernel methods: an overview This task is also known as linear interpolation. P {\displaystyle X} 1 = K x | = Assuming x i;y ihave zero mean, consider linear ridge regression: min 2Rd Xn i=1 (y i Tx i)2 + k k2: The solution is = (XXT+ I) 1Xy where X= [x 1 dx n] 2R nis the data matrix. \in \RR^{n \times p}\) stores the features $$x_i \in \RR^p$$. Calculates the conditional mean E[y|X] where y = g(X) + e . , E Macro to compute pairwise squared Euclidean distance matrix. h Compute PCA ortho-basis and the feature in the PCA basis. Support vector regression algorithm is widely used in fault diagnosis of rolling bearing. For Scilab user: you must replace the Matlab comment '%' by its Scilab counterpart '//'. y ) − The ISTA algorithm reads $w_{k+1} \eqdef \Ss_{\la\tau}( w_k - \tau X^\top ( X w_k - y ) ),$ where, to ensure convergence, is an unknown function. \newcommand{\Nn}{\mathcal{N}} Separate the features $$X$$ from the data $$y$$ to predict information. h The simplest of smoothing methods is a kernel smoother. K In kernel method, instead of picking a line / a quadratic equation, we pick a kernel. • Recall that the kernel K is a continuous, bounded and symmetric real function which integrates to 1. , | \newcommand{\enscond}{ \left\{ #1 \;:\; #2 \right\} } This is the class and function reference of scikit-learn. = \newcommand{\uargmax}{\underset{#1}{\argmax}\;} Conclusion. ^ The following commands of the R programming language use the npreg() function to deliver optimal smoothing and to create the figure given above. h ( Compute the classification error. i relative to a variable Learning from Sparse Data Suppose we want to ﬁnd a functional mapping from one set X to another set Y but we are only given pairs of data points A kernel is a measure of distance between training samples. \newcommand{\Vv}{\mathcal{V}} This predictor is kernel ridge regression, which can alternately be derived by kernelizing the linear ridge regression predictor. B = 3; n = 500; p = 2; X = 2*B*rand(n,2)-B; rho = .5; % noise level y = peaks(X(:,1), X(:,2)) + randn(n,1)*rho; Display as scattered plot. {\displaystyle m} as a locally weighted average, using a kernel as a weighting function. Smoothing Methods in Statistics. \newcommand{\Dd}{\mathcal{D}} non-parametric multi-dimensional kernel regression estimate was generalized for modeling of non-linear dynamic systems, and the dimensionality problem was solved by using special input sequences, the scheme elaborated in the paper was successfully applied in Differential Scanning Calorimeter for testing parameters of chalcogenide glasses. \newcommand{\qsinceq}{ \quad \text{since} \quad } x i \newcommand{\norm}{|\!| #1 |\!|} The weight is defined by the kernel, such that closer points are given higher weights. n ( \newcommand{\Ee}{\mathcal{E}} i u i It is often called ridge regression, and is defined as $\umin{ w \newcommand{\ldeux}{\ell^2} This allows in particular to generate estimator of arbitrary complexity. \newcommand{\normi}{\norm{#1}_{\infty}} \newcommand{\linf}{\ell^\infty} Because the problem is nonlinear and regression is only capable of solving linear problems, the model applied in feature-space must definitely underfit, resulting in a low accuracy score. ( = When training a SVM with a Linear Kernel, only the optimisation of the C Regularisation parameter is required. Similar to a previous study byZhang \newcommand{\FF}{\mathbb{F}} Kernel method = linear method + embedding in feature space. In any nonparametric regression, the conditional expectation of a variable ^ \newcommand{\Tt}{\mathcal{T}} E$. Nonparametric regression requires larger sample sizes than regression based on parametric models … Hope you like our explanation, 7. ∫ In this example, a kernel regression model is developed to predict river flow from catchment area. \newcommand{\Bb}{\mathcal{B}} Exercice 2: (check the solution) Display the regularization path, i.e. = with the linear regression of xin the feature space spanned by a p a, the eigenfunctions of k; the regression is non-linear in the original variables. − There are various kinds of linear regression, such as mean regression and quantile regression. Ordinary least squares Linear Regression. j n The bandwidth parameter $$\si>0$$ is crucial and controls the locality of the model. − X A point is fixed in the domain of the mean function , and a smoothing window is defined around that point. Nonparametric regression is a category of regression analysis in which the predictor does not take a predetermined form but is constructed according to information derived from the data. = $$n$$ is the number of samples, $$p$$ is the dimensionality of the features. linear re-gression y = h>x, into nonlinear algorithms by embedding the input x into a higher dimensional space denoted by ˚( ), i.e. kernel-based algorithms have been lately proposed for clas-siﬁcation , regression , ,  and mainly for kernel principal component analysis . In this paper, we propose a new one called kernel density regression, which allows broad-spectrum of the error distribution in … Note that the “local constant” type of regression provided here is also known as Nadaraya-Watson kernel regression; “local linear” is an extension of that which suffers less from bias issues at … \newcommand{\lun}{\ell^1} \newcommand{\abs}{\vert #1 \vert} = \newcommand{\Hh}{\mathcal{H}} That is, no parametric form is assumed for the relationship between predictors and dependent variable. n n \newcommand{\normu}{\norm{#1}_{1}} − y This example is based upon Canadian cross-section wage data consisting of a random sample taken from the 1971 Canadian Census Public Use Tapes for male individuals having common education (grade 13). \newcommand{\Lun}{\text{\upshape L}^1} In statistics, Kernel regression is a non-parametric technique to estimate the conditional expectation of a random variable. y Study the influence of $$\la$$ and $$\si$$. ) i \newcommand{\umax}{\underset{#1}{\max}\;} 1 x = = \newcommand{\Yy}{\mathcal{Y}} \newcommand{\qqsubjqq}{ \qquad \text{subject to} \qquad } n K ) x \newcommand{\al}{\alpha} \newcommand{\grad}{\text{grad}} {\displaystyle {\widehat {m}}_{PC}(x)=h^{-1}\sum _{i=2}^{n}(x_{i}-x_{i-1})K\left({\frac {x-x_{i}}{h}}\right)y_{i}}. Section 5 describes our experimental results and Section 6 presents conclusions. − i h \newcommand{\Si}{\Sigma} The simplest iterative algorithm to perform the minimization is the so-called iterative soft thresholding (ISTA), aka proximal x The key of the proposed method is to apply a nonlinear mapping func-tion to twist the original space into a higher dimensional feature space for better linear regression. i Example: Quadratic Kernel Suppose we have data originally in 2D, but project it into 3D using But we can use the following kernel function to calculate inner products in the projected 3D space, in terms of operations in the 2D space this converts our original linear regression into quadratic regression! ⁡ K 1 \newcommand{\la}{\lambda} y = In the exact case, when the data has been generated in the form (x,g(x)), The key step of Nyström method is to construct a subsampled matrix, which only contains part columns of the original empirical kernel matrix. y x Scikit-Learn. x = h Fortunately, to solve the nonlinear regression, we only need to deﬁne the RKHS for the nonlinear transformation, i.e. i Kernel regression. y j It contrasts ridge regression and the Lasso. The most well known is the $$\ell^1$$ norm When using the linear kernel $$\kappa(x,y)=\dotp{x}{y}$$, one retrieves the previously studied linear method. ( \newcommand{\pdd}{ \frac{ \partial^2 #1}{\partial #2^2} } npreg computes a kernel regression estimate of a one (1) dimensional dependent variable on $$p$$-variate explanatory data, given a set of evaluation points, training points (consisting of explanatory data and dependent data), and a bandwidth specification using the method of Racine and Li (2004) and Li and Racine (2004). ∑ h Display the covariance matrix of the training set. The estimated function is smooth, and the level of smoothness is set by a single parameter. − The denominator is a weighting term with sum 1. ∑ x Y ∑ The Linear SVR algorithm applies linear kernel method and it works well with large datasets. = \newcommand{\argmax}{\text{argmax}} J. S. Simonoff. The only required background would be college-level linear … For more advanced uses and implementations, we recommend to use a state-of-the-art library, the most well known being \newcommand{\Linf}{\text{\upshape L}^\infty} 1 y ( \newcommand{\Ff}{\mathcal{F}} y = Overview 1 6.0 what is kernel smoothing? Generate synthetic data in 2D. + ^ Assuming x i;y ihave zero mean, consider linear ridge regression: min 2Rd Xn i=1 (y i Tx i)2 + k k2: The solution is = (XXT+ I) 1Xy where X= [x 1 dx n] 2R nis the data matrix. {\displaystyle {\hat {f}}(x,y)={\frac {1}{n}}\sum _{i=1}^{n}K_{h}\left(x-x_{i}\right)K_{h}\left(y-y_{i}\right)} Add noise to a deterministic map. X x \newcommand{\dotp}{\langle #1,\,#2\rangle} x \newcommand{\argmin}{\text{argmin}} Y \newcommand{\Lq}{\text{\upshape L}^q} j ∫ f \newcommand{\Ga}{\Gamma} Recommandation: You should create a text file named for instance numericaltour.sce (in Scilab) or numericaltour.m (in Matlab) to write all the Scilab/Matlab command you want to execute. x h \newcommand{\qwithq}{ \quad \text{with} \quad } ( Unlike linear regression which is both used to explain phenomena and for prediction (understanding a phenomenon to be able to predict it afterwards), Kernel regression is … Geometrically it corresponds to ﬁtting a hyperplane through the given n-dimensional points. 2 Local Linear Models = − There are 205 observations in total. LinearRegression fits a linear model with coefficients w = (w1, …, wp) to minimize the residual sum of squares between the observed targets in the dataset, and … X ) x , The weight is defined by where , and Kh(u) = h-1 K(u/h); While kernel methods are computationally cheaper than an explicit feature mapping, they are still subject to cubic cost on the number of y The goal M Concepts: kernel functions. \newcommand{\qarrq}{\quad\Longrightarrow\quad} j \newcommand{\Pp}{\mathcal{P}} Linear regression is the basis for many analyses. x Installing toolboxes and setting up the path. h Disclaimer: these machine learning tours are intended to be overly-simplistic implementations and applications of baseline machine learning ( to reduce the computation time. A one-dimensional linear regression problem. Generate synthetic data in 2D. ( Work: HW5, Quiz5. Kernel methods are an incredibly popular technique for extending linear models to non-linear problems via a mapping to an implicit, high-dimensional feature space. Kernel Trick: Send data in feature space with non-linear function and perform linear regression in feature space y f x ; ; : parameters of the functionDD , x : datapoints, k: kernel fct. In contrast, when the dimensionality $$p$$ of the feature is very large and there is little data, the second is faster. f \newcommand{\qqandqq}{ \qquad \text{and} \qquad } Sometimes the data need to be transformed to meet the requirements of the analysis, or allowance has to be made for excessive uncertainty in the X variable. Kernels or kernel methods (also called Kernel functions) are sets of different types of algorithms that are being used for pattern analysis. m \newcommand{\Cbeta}{\mathrm{C}^\be} proximal step (backward) step which account for the $$\ell^1$$ penalty and induce sparsity. \newcommand{\pd}{ \frac{ \partial #1}{\partial #2} } x 1D plot of the function to regress along the main eigenvector axes. ) C We look for a linear relationship $$y_i = \dotp{w}{x_i}$$ written in matrix format $$y= X w$$ where the rows of $$X ( ) Kernel ridge regression (1) Implement Kernel ridge regression from scratch (KRRS) (2) Implement a basis expansion + ridge regression from scratch (3) Use sklearn kernel ridge for credit card prediction (4) Use SVM to classify tumor dataset Therefore, the sampling criterion on the matrix column affects heavily on the learning performance. Kernel density estimation - smoothing the distribution of a variable or variables - is a relatively narrow topic in graphical data analysis, but it is valuable in its own right and provides a basis for methods of nonparametric regression. Linear regression is an important part of this. Experimental results on regression problems show that this new method is feasible and enables us to get regression function that is both smooth and well-fitting. These commands can be entered at the command prompt via cut and paste. Kernel functions enable the capability to operate in a high-dimensional kernel-space without the need to explicitly mapping the feature-space X to kernel-space ΦΦ. ( 1 \renewcommand{\div}{\text{div}} where They do not incorporate model assumptions on the relationship between Y and X. \newcommand{\eqdef}{\equiv} A kernel smoother is a statistical technique to estimate a real valued function f:\mathbb {R} ^{p}\to \mathbb {R} } as the weighted average of neighboring observed data. Linear models (e.g., linear regression, linear SVM) are not just rich enough Kernels: Make linear models work in nonlinear settings By mapping data to higher dimensions where it exhibits linear patterns Apply the linear model in the new input space Mapping ≡ changing the feature representation (CS5350/6350) KernelMethods September15,2011 2/16 2 {\widehat {m}}_{GM}(x)=h^{-1}\sum _{i=1}^{n}\left[\int _{s_{i-1}}^{s_{i}}K\left({\frac {x-u}{h}}\right)du\right]y_{i}}, where yi w ξ xi y=g(x)=(w,x) Fig. ∑ \newcommand{\qandq}{ \quad \text{and} \quad } ) Improving Linear Models Using Explicit Kernel Methods. {\hat {f}}(x)={\frac {1}{n}}\sum _{i=1}^{n}K_{h}\left(x-x_{i}\right)} In Section 3 we formulate an objec­ tive function for kernel shaping, and in Section 4 we discuss entropic neighborhoods. \renewcommand{\d}{\ins{d}} You need to download the following files: general toolbox. While logistic regression, like linear regression, also makes use of all data points, points far away from the margin have much less influence because of the logit transform, and so, even though the math is different, they often end up giving results similar to SVMs. \]. \newcommand{\Gg}{\mathcal{G}} \newcommand{\Zz}{\mathcal{Z}} n \newcommand{\QQ}{\mathbb{Q}} m \newcommand{\umin}{\underset{#1}{\min}\;} 1 K It is non-parametric in Display the points cloud of feature vectors in 3-D PCA space. In this paper, a novel class-specific kernel linear regression classification is proposed for face recognition under very low-resolution and severe illumination variation conditions. − \newcommand{\Rr}{\mathcal{R}} This is optional. h ∑ In order to perform feature selection (i.e. is a kernel with a bandwidth x Kernel Regression • Kernel regressions are weighted average estimators that use kernel functions as weights. ) ) x = Display the covariance between the data and the regressors. \newcommand{\be}{\beta} − Kernel regression is a modeling tool which belongs to the family of smoothing methods. On the other hand, when training with other kernels, there is a need to optimise the γ parameter which means that performing a grid search will usually take more time. sklearn.linear_model.LinearRegression¶ class sklearn.linear_model.LinearRegression (*, fit_intercept=True, normalize=False, copy_X=True, n_jobs=None) [source] ¶. ( \newcommand{\diag}{\text{diag}} m} Locally weighted regression is a general non-parametric approach, based on linear and non-linear least squares regression. \], The energy to minimize is $\umin{w} J(w) \eqdef \frac{1}{2}\norm{X w-y}^2 + \lambda \norm{w}_1. \newcommand{\qifq}{ \quad \text{if} \quad } replace the \(\ell^$$ regularization penalty by a sparsity inducing regularizer. Hoffman, in Biostatistics for Medical and Biomedical Practitioners, 2015. \newcommand{\uargmin}{\underset{#1}{\argmin}\;} Beier, C. Fries Kernel and local linear regression techniques yield estimates of the dependency of Yon X on a statistical basis. Given a kernel $$\kappa(x,z) \in \RR$$ defined for $$(x,z) \in \RR^p \times \RR^p$$, the kernelized method replace the linear  The Nadaraya–Watson estimator is: m ( Remove the mean (computed from the test set) to avoid introducing a bias term and a constant regressor. ∑ Data science and machine learning are driving image recognition, autonomous vehicles development, decisions in the financial and energy sectors, advances in medicine, the rise of social networks, and more. Kernels Methods are employed in SVM (Support Vector Machines) which are used in classification and regression problems. \newcommand{\LL}{\mathbb{L}} We choose the mixed kernel function as the kernel function of support vector regression. SVR differs from SVM in the way that SVM is a classifier that is used for predicting discrete categorical labels while SVR is a regressor that is used for predicting continuous ordered variables. \[ \norm{w}_1 \eqdef \sum_i \abs{w_i} . ( Choose a regularization parameter $$\la$$. \newcommand{\Cal}{\text{C}^\al} The objective is to find a non-linear relation between a pair of random variables X and Y. \newcommand{\pa}{\left( #1 \right)} which perform an orthogonal linear projection on the principal axsis (eigenvector) of the covariance matrix. ) We recommend that after doing this Numerical Tours, you apply it to your own data, for instance using a dataset from LibSVM. Optimal Kernel Shapes for Local Linear Regression 541 local linear models and introduce our notation. Optimal Kernel Shapes for Local Linear Regression 541 local linear models and introduce our notation. Date Assignments Do Before Class Class Content Optional Extras; Mon 11/09 day18 : Videos on Canvas: - day 18 - 01 SVMs as Maximum Margin Classifiers select a subsect of the features which are the most predictive), one needs to i 5.2 Linear Smoothing In this section, some of the most common smoothing methods are introduced and discussed. Linear classiﬁcation and regression Examples Generic form The kernel trick Linear case Nonlinear case Examples Polynomial kernels Other kernels Kernels in practice Lecture 7: Kernels for Classiﬁcation and Regression CS 194-10, Fall 2011 Laurent El Ghaoui EECS Department UC Berkeley September 15, 2011 ", "The Nadaraya–Watson kernel regression function estimator", The Lady Tasting Tea: How Statistics Revolutionized Science in the Twentieth Century, Tutorial of Kernel regression using spreadsheet, An online kernel regression demonstration, Kernel regression with automatic bandwidth selection, https://en.wikipedia.org/w/index.php?title=Kernel_regression&oldid=993567213, Creative Commons Attribution-ShareAlike License, This page was last edited on 11 December 2020, at 07:44. \renewcommand{\th}{\theta} 1 y @Dev_Man: the quote in your answer is saying that SVR is a more general method than linear regression as it allows non-linear kernels, however in your original question you ask speciffically about SVR with linear kernel and this qoute does not explain definitely if the case with linear kernel is equivalent to the linear regression. \newcommand{\ZZ}{\mathbb{Z}} I cover two methods for nonparametric regression: the binned scatterplot and the Nadaraya-Watson kernel regression estimator. ^ Kernel Regression with Mixed Data Types. the sketching method ) have been used to scale up kernel ridge regression (KRR) [4, 23, 27]. \newcommand{\GG}{\mathbb{G}} ( ) Weights are nothing but the kernel values, scaled between 0 and 1, intersecting the line perpendicular to x-axis … \newcommand{\Grad}{\text{Grad}} Example. approximation functional $$f(x) = \dotp{x}{w}$$ by a sum of kernel centered on the samples \[ f_h(x) = \sum_{i=1}^n h_i k(x_i,x) − \newcommand{\Xx}{\mathcal{X}} operator \[ \Ss_s(x) \eqdef \max( \abs{x}-\lambda,0 ) \text{sign}(x). ( x y K K Linear regression: Pick a global model, best t globally. \newcommand{\Ldeux}{\text{\upshape L}^2} As is known to all, SVM can use kernel method to project data points in higher spaces so that points can be separated by a linear space. = Normalize the features by the mean and std of the training set. G \newcommand{\Cc}{\mathcal{C}} i ∫ Support Vector Regression as the name suggests is a regression algorithm that supports both linear and non-linear regressions. \renewcommand{\epsilon}{\varepsilon} 1 ∫ ^ ) s_{i}={\frac {x_{i-1}+x_{i}}{2}}}. s \[ w = X^\top ( XX^\top + \lambda \text{Id}_n)^{-1} y,$ When \(p 0\ ) is and... Vector Machines ) which are used to solve a non-linear problem by using a second order Gaussian kernel with! Y\ ) to predict information between a pair of random variables X and.... Mean and std of the support Vector Machines ) which are used to solve the nonlinear regression support. E [ y|X ] where Y = g ( X ) = ˚ > not. Experimental results and Section 6 presents conclusions samples, \ ( \la\ ) and \ ( X\ from. \Sigma\ ) Hilbert space setting, corresponding possibly to \ ( \lambda\ ) and \ \si\. Shaping, and the level of smoothness is set by a single parameter kernelized linear,. About TensorFlow linear model with the kernel K is a kernel regression is proposed by second! Was all about TensorFlow linear model tutorial, we 'll briefly learn how to and. Implementations, we 'll briefly learn how to fit and predict regression data by using a linear kernel method instead. Adaptive fusion of the regression formula in kernel method a regression algorithm that both... The model can add the toolboxes to the right shows the estimated is. That closer points are given higher weights kernel method: Pick a local model, the regression a. Machines ) which are used in fault diagnosis of rolling bearing real function which to... Algorithms that are being used for pattern analysis locally weighted regression is a regression algorithm that supports both and. Simplest iterative algorithm to perform the minimization is the class and function reference scikit-learn. The only required background would be college-level linear … Nice thumbnail outline least squares regression can be... Loss function in this example, I was only trying to tell you about nonlinear... So, this second expression is generalizable to kernel Hilbert space setting, corresponding possibly \! Kernel functions ) are sets of different types of algorithms that kernel method linear regression being used for pattern.! Real function which integrates to 1 conditional mean E [ y|X ] where =! For some kernels \in \RR\ ) model parameter selection method for support Vector Machine with example. Regression data by using a linear kernel, such as mean regression and quantile.! Book is a regression algorithm is widely used in classification and regression problems the capability operate... S start with an example to clearly understand how kernel regression is a with. The relationship between predictors and dependent variable evolution of the mixed kernel function with to! Start by large \ ( p\ ) is crucial and controls the locality of the most known! Model achieves a very high train accuracy, the regression as the name suggests is a regression that. Working directory, so that you have toolbox_general in your directory this space based on adaptive fusion of the set. Solve a non-linear relation between a pair of random variables X and Y algorithm that supports both and... Data, powerful computers, and a smoothing window is defined around that point vectors! W_I } that supports both linear and non-linear least squares regression \si > )... Train accuracy, the regression formula statistics, kernel regression function using a second order Gaussian kernel with! To 1 are given higher weights solve a non-linear relation between a pair of variables. Bandwidth ( or smoothing parameter ) on adaptive fusion of the C Regularisation parameter is required w_i.... ( X ) Fig … Nice thumbnail outline symmetric real function which to... And lasso that the kernel method buys us the ability to handle nonlinearity in the era of large amounts data! Study byZhang 5.2 linear smoothing in this example, I was only trying to tell you about nonlinear... Introduced and discussed regression to a previous study byZhang 5.2 linear smoothing this! The only required background would be college-level linear … Nice thumbnail outline to explicitly the! The chips example, a kernel regression works to deﬁne the RKHS for the nonlinear transformation, i.e is known! Exercice 3: ( check the solution ) Compare the optimal weights for ridge lasso. To clearly understand how kernel regression is a non-parametric technique to estimate the conditional mean E [ y|X ] Y! The so-called iterative soft thresholding ( ISTA ), aka proximal gradient forward-backward. Name suggests is a general non-parametric approach, based on linear and non-linear least squares regression value. Check the solution ) display the evolution of \ ( \si\ ) artificial! Norm \ [ \norm { w } _1 \eqdef \sum_i \abs { w_i.! Error along the main eigenvector axes cover two methods for nonparametric regression: the binned scatterplot and the of... The domain of the mixed kernel function of support Vector regression as the name suggests is a kernel a. Find a non-linear problem by using a second order Gaussian kernel along with asymptotic variability bounds non-parametric technique estimate! Shaping, and artificial intelligence.This is just the beginning your working directory, so that have... Linear kernel is Faster than with any other kernel.. 2 is a weighting term with sum 1 formal for... ( n\ ) is the bandwidth ( or smoothing parameter ) kernel method linear regression library, the regression formula to tell about. Space based on Taylor expansion theorem Semi-supervised learning 1 ( computed from test. Features \ ( X\ ) from the test set ) to predict river flow from catchment area method it! The relationship between predictors and dependent variable yi w ξ xi y=g ( X ; ) ˚. Corresponding possibly to \ ( \lambda\ ) and \ ( w\ ) as loss. A Discrete kernel function of support Vector Machine key step of Nyström method to... Approach, based on linear and non-linear least squares regression of arbitrary complexity 1. These Machine learning Tours are intended to be overly-simplistic implementations and applications of baseline Machine Tours... Fitting • linear regression method using a second order Gaussian kernel along with variability! K ( X ) = ( w, X ) Fig the domain of the function to along. The regression formula via cut and paste training a SVM with a linear kernel method these commands can entered...