statsmodels ols multiple regression

Create a Model from a formula and dataframe. FYI, note the import above. Replacing broken pins/legs on a DIP IC package, AC Op-amp integrator with DC Gain Control in LTspice. Explore the 10 popular blogs that help data scientists drive better data decisions. Introduction to Linear Regression Analysis. 2nd. Just pass. All other measures can be accessed as follows: Step 1: Create an OLS instance by passing data to the class m = ols (y,x,y_varnm = 'y',x_varnm = ['x1','x2','x3','x4']) Step 2: Get specific metrics To print the coefficients: >>> print m.b To print the coefficients p-values: >>> print m.p """ y = [29.4, 29.9, 31.4, 32.8, 33.6, 34.6, 35.5, 36.3, from_formula(formula,data[,subset,drop_cols]). R-squared: 0.353, Method: Least Squares F-statistic: 6.646, Date: Wed, 02 Nov 2022 Prob (F-statistic): 0.00157, Time: 17:12:47 Log-Likelihood: -12.978, No. Bursts of code to power through your day. Then fit () method is called on this object for fitting the regression line to the data. Did any DOS compatibility layers exist for any UNIX-like systems before DOS started to become outmoded? I also had this problem as well and have lots of columns needed to be treated as categorical, and this makes it quite annoying to deal with dummify. OLSResults (model, params, normalized_cov_params = None, scale = 1.0, cov_type = 'nonrobust', cov_kwds = None, use_t = None, ** kwargs) [source] Results class for for an OLS model. Indicates whether the RHS includes a user-supplied constant. Multiple regression - python - statsmodels, Catch multiple exceptions in one line (except block), Create a Pandas Dataframe by appending one row at a time, Selecting multiple columns in a Pandas dataframe. endog is y and exog is x, those are the names used in statsmodels for the independent and the explanatory variables. http://statsmodels.sourceforge.net/stable/generated/statsmodels.regression.linear_model.RegressionResults.predict.html with missing docstring, Note: this has been changed in the development version (backwards compatible), that can take advantage of "formula" information in predict If none, no nan Parameters: endog array_like. Thanks for contributing an answer to Stack Overflow! estimation by ordinary least squares (OLS), weighted least squares (WLS), Additional step for statsmodels Multiple Regression? Hear how DataRobot is helping customers drive business value with new and exciting capabilities in our AI Platform and AI Service Packages. I divided my data to train and test (half each), and then I would like to predict values for the 2nd half of the labels. We would like to be able to handle them naturally. Web[docs]class_MultivariateOLS(Model):"""Multivariate linear model via least squaresParameters----------endog : array_likeDependent variables. predictions = result.get_prediction (out_of_sample_df) predictions.summary_frame (alpha=0.05) I found the summary_frame () method buried here and you can find the get_prediction () method here. Copyright 2009-2019, Josef Perktold, Skipper Seabold, Jonathan Taylor, statsmodels-developers. OLS has a With a goal to help data science teams learn about the application of AI and ML, DataRobot shares helpful, educational blogs based on work with the worlds most strategic companies. In Ordinary Least Squares Regression with a single variable we described the relationship between the predictor and the response with a straight line. WebThe first step is to normalize the independent variables to have unit length: [22]: norm_x = X.values for i, name in enumerate(X): if name == "const": continue norm_x[:, i] = X[name] / np.linalg.norm(X[name]) norm_xtx = np.dot(norm_x.T, norm_x) Then, we take the square root of the ratio of the biggest to the smallest eigen values. All rights reserved. What am I doing wrong here in the PlotLegends specification? In the following example we will use the advertising dataset which consists of the sales of products and their advertising budget in three different media TV, radio, newspaper. The selling price is the dependent variable. Not the answer you're looking for? Webstatsmodels.multivariate.multivariate_ols._MultivariateOLS class statsmodels.multivariate.multivariate_ols._MultivariateOLS(endog, exog, missing='none', hasconst=None, **kwargs)[source] Multivariate linear model via least squares Parameters: endog array_like Dependent variables. we let the slope be different for the two categories. It means that the degree of variance in Y variable is explained by X variables, Adj Rsq value is also good although it penalizes predictors more than Rsq, After looking at the p values we can see that newspaper is not a significant X variable since p value is greater than 0.05. For more information on the supported formulas see the documentation of patsy, used by statsmodels to parse the formula. Why does Mister Mxyzptlk need to have a weakness in the comics? A nobs x k_endog array where nobs isthe number of observations and k_endog is the number of dependentvariablesexog : array_likeIndependent variables. D.C. Montgomery and E.A. Lets do that: Now, we have a new dataset where Date column is converted into numerical format. We have no confidence that our data are all good or all wrong. Why do small African island nations perform better than African continental nations, considering democracy and human development? A 1-d endogenous response variable. Does Counterspell prevent from any further spells being cast on a given turn? This is part of a series of blog posts showing how to do common statistical learning techniques with Python. Class to hold results from fitting a recursive least squares model. Recovering from a blunder I made while emailing a professor. Asking for help, clarification, or responding to other answers. If you would take test data in OLS model, you should have same results and lower value Share Cite Improve this answer Follow GLS is the superclass of the other regression classes except for RecursiveLS, How can I check before my flight that the cloud separation requirements in VFR flight rules are met? Web[docs]class_MultivariateOLS(Model):"""Multivariate linear model via least squaresParameters----------endog : array_likeDependent variables. The OLS () function of the statsmodels.api module is used to perform OLS regression. To subscribe to this RSS feed, copy and paste this URL into your RSS reader. Connect and share knowledge within a single location that is structured and easy to search. OLSResults (model, params, normalized_cov_params = None, scale = 1.0, cov_type = 'nonrobust', cov_kwds = None, use_t = None, ** kwargs) [source] Results class for for an OLS model. The summary () method is used to obtain a table which gives an extensive description about the regression results Syntax : statsmodels.api.OLS (y, x) A 50/50 split is generally a bad idea though. Python sort out columns in DataFrame for OLS regression. However, once you convert the DataFrame to a NumPy array, you get an object dtype (NumPy arrays are one uniform type as a whole). hessian_factor(params[,scale,observed]). Full text of the 'Sri Mahalakshmi Dhyanam & Stotram'. Browse other questions tagged, Where developers & technologists share private knowledge with coworkers, Reach developers & technologists worldwide. rev2023.3.3.43278. For the Nozomi from Shinagawa to Osaka, say on a Saturday afternoon, would tickets/seats typically be available - or would you need to book? It should be similar to what has been discussed here. predictions = result.get_prediction (out_of_sample_df) predictions.summary_frame (alpha=0.05) I found the summary_frame () method buried here and you can find the get_prediction () method here. ratings, and data applied against a documented methodology; they neither represent the views of, nor Can I do anova with only one replication? We can clearly see that the relationship between medv and lstat is non-linear: the blue (straight) line is a poor fit; a better fit can be obtained by including higher order terms. Copyright 2009-2019, Josef Perktold, Skipper Seabold, Jonathan Taylor, statsmodels-developers. Is it possible to rotate a window 90 degrees if it has the same length and width? (R^2) is a measure of how well the model fits the data: a value of one means the model fits the data perfectly while a value of zero means the model fails to explain anything about the data. https://www.statsmodels.org/stable/example_formulas.html#categorical-variables. Find centralized, trusted content and collaborate around the technologies you use most. Browse other questions tagged, Where developers & technologists share private knowledge with coworkers, Reach developers & technologists worldwide. Often in statistical learning and data analysis we encounter variables that are not quantitative. If you would take test data in OLS model, you should have same results and lower value Share Cite Improve this answer Follow Not the answer you're looking for? Application and Interpretation with OLS Statsmodels | by Buse Gngr | Analytics Vidhya | Medium Write Sign up Sign In 500 Apologies, but something went wrong on our end. Do roots of these polynomials approach the negative of the Euler-Mascheroni constant? MacKinnon. A p x p array equal to \((X^{T}\Sigma^{-1}X)^{-1}\). Multiple Linear Regression: Sklearn and Statsmodels | by Subarna Lamsal | codeburst 500 Apologies, but something went wrong on our end. By clicking Post Your Answer, you agree to our terms of service, privacy policy and cookie policy. To subscribe to this RSS feed, copy and paste this URL into your RSS reader. Read more. There are missing values in different columns for different rows, and I keep getting the error message: These are the next steps: Didnt receive the email? OLS (endog, exog = None, missing = 'none', hasconst = None, ** kwargs) [source] Ordinary Least Squares. The variable famhist holds if the patient has a family history of coronary artery disease. There are several possible approaches to encode categorical values, and statsmodels has built-in support for many of them. What can a lawyer do if the client wants him to be acquitted of everything despite serious evidence? To learn more, see our tips on writing great answers. You're on the right path with converting to a Categorical dtype. Full text of the 'Sri Mahalakshmi Dhyanam & Stotram'. This is equal n - p where n is the rev2023.3.3.43278. Greene also points out that dropping a single observation can have a dramatic effect on the coefficient estimates: We can also look at formal statistics for this such as the DFBETAS a standardized measure of how much each coefficient changes when that observation is left out. With the LinearRegression model you are using training data to fit and test data to predict, therefore different results in R2 scores. The code below creates the three dimensional hyperplane plot in the first section. see http://statsmodels.sourceforge.net/stable/generated/statsmodels.regression.linear_model.OLS.predict.html. Evaluate the Hessian function at a given point. What can a lawyer do if the client wants him to be acquitted of everything despite serious evidence? Is the God of a monotheism necessarily omnipotent? This class summarizes the fit of a linear regression model. Can Martian regolith be easily melted with microwaves? All other measures can be accessed as follows: Step 1: Create an OLS instance by passing data to the class m = ols (y,x,y_varnm = 'y',x_varnm = ['x1','x2','x3','x4']) Step 2: Get specific metrics To print the coefficients: >>> print m.b To print the coefficients p-values: >>> print m.p """ y = [29.4, 29.9, 31.4, 32.8, 33.6, 34.6, 35.5, 36.3, Statsmodels is a Python module that provides classes and functions for the estimation of different statistical models, as well as different statistical tests. If raise, an error is raised. What is the naming convention in Python for variable and function? Using categorical variables in statsmodels OLS class. Lets take the advertising dataset from Kaggle for this. These (R^2) values have a major flaw, however, in that they rely exclusively on the same data that was used to train the model. Lets say youre trying to figure out how much an automobile will sell for. That is, the exogenous predictors are highly correlated. WebI'm trying to run a multiple OLS regression using statsmodels and a pandas dataframe. This is because 'industry' is categorial variable, but OLS expects numbers (this could be seen from its source code). And I get, Using categorical variables in statsmodels OLS class, https://www.statsmodels.org/stable/example_formulas.html#categorical-variables, statsmodels.org/stable/examples/notebooks/generated/, How Intuit democratizes AI development across teams through reusability. If so, how close was it? Subarna Lamsal 20 Followers A guy building a better world. Click the confirmation link to approve your consent. Multiple Linear Regression: Sklearn and Statsmodels | by Subarna Lamsal | codeburst 500 Apologies, but something went wrong on our end. autocorrelated AR(p) errors. We first describe Multiple Regression in an intuitive way by moving from a straight line in a single predictor case to a 2d plane in the case of two predictors. Notice that the two lines are parallel. Predicting values using an OLS model with statsmodels, http://statsmodels.sourceforge.net/stable/generated/statsmodels.regression.linear_model.OLS.predict.html, http://statsmodels.sourceforge.net/stable/generated/statsmodels.regression.linear_model.RegressionResults.predict.html, http://statsmodels.sourceforge.net/devel/generated/statsmodels.regression.linear_model.RegressionResults.predict.html, How Intuit democratizes AI development across teams through reusability. changing the values of the diagonal of a matrix in numpy, Statsmodels OLS Regression: Log-likelihood, uses and interpretation, Create new column based on values from other columns / apply a function of multiple columns, row-wise in Pandas, The difference between the phonemes /p/ and /b/ in Japanese. I want to use statsmodels OLS class to create a multiple regression model. model = OLS (labels [:half], data [:half]) predictions = model.predict (data [half:]) Consider the following dataset: I've tried converting the industry variable to categorical, but I still get an error. Depending on the properties of \(\Sigma\), we have currently four classes available: GLS : generalized least squares for arbitrary covariance \(\Sigma\), OLS : ordinary least squares for i.i.d. Despite its name, linear regression can be used to fit non-linear functions. Whats the grammar of "For those whose stories they are"? Hence the estimated percentage with chronic heart disease when famhist == present is 0.2370 + 0.2630 = 0.5000 and the estimated percentage with chronic heart disease when famhist == absent is 0.2370. The simplest way to encode categoricals is dummy-encoding which encodes a k-level categorical variable into k-1 binary variables. The final section of the post investigates basic extensions. I divided my data to train and test (half each), and then I would like to predict values for the 2nd half of the labels. The nature of simulating nature: A Q&A with IBM Quantum researcher Dr. Jamie We've added a "Necessary cookies only" option to the cookie consent popup. How can this new ban on drag possibly be considered constitutional? Why do many companies reject expired SSL certificates as bugs in bug bounties? Also, if your multivariate data are actually balanced repeated measures of the same thing, it might be better to use a form of repeated measure regression, like GEE, mixed linear models , or QIF, all of which Statsmodels has. RollingWLS(endog,exog[,window,weights,]), RollingOLS(endog,exog[,window,min_nobs,]). ProcessMLE(endog,exog,exog_scale,[,cov]). A regression only works if both have the same number of observations. The Python code to generate the 3-d plot can be found in the appendix. Thanks for contributing an answer to Stack Overflow! Evaluate the score function at a given point. I want to use statsmodels OLS class to create a multiple regression model. The OLS () function of the statsmodels.api module is used to perform OLS regression. How can I access environment variables in Python? Empowering Kroger/84.51s Data Scientists with DataRobot, Feature Discovery Integration with Snowflake, DataRobot is committed to protecting your privacy. Asking for help, clarification, or responding to other answers. Minimising the environmental effects of my dyson brain, Using indicator constraint with two variables. Splitting data 50:50 is like Schrodingers cat. If you had done: you would have had a list of 10 items, starting at 0, and ending with 9. Site design / logo 2023 Stack Exchange Inc; user contributions licensed under CC BY-SA. and can be used in a similar fashion. The n x n upper triangular matrix \(\Psi^{T}\) that satisfies Parameters: Using statsmodel I would generally the following code to obtain the roots of nx1 x and y array: But this does not work when x is not equivalent to y. Just another example from a similar case for categorical variables, which gives correct result compared to a statistics course given in R (Hanken, Finland). Do roots of these polynomials approach the negative of the Euler-Mascheroni constant? The value of the likelihood function of the fitted model. This is because the categorical variable affects only the intercept and not the slope (which is a function of logincome). Is it possible to rotate a window 90 degrees if it has the same length and width? We can show this for two predictor variables in a three dimensional plot. It returns an OLS object. They are as follows: Now, well use a sample data set to create a Multiple Linear Regression Model. Browse other questions tagged, Where developers & technologists share private knowledge with coworkers, Reach developers & technologists worldwide, @user333700 Even if you reverse it around it has the same problems of a nx1 array. exog array_like Is a PhD visitor considered as a visiting scholar? Now, its time to perform Linear regression. Thanks for contributing an answer to Stack Overflow! Share Improve this answer Follow answered Jan 20, 2014 at 15:22 The dependent variable. These are the different factors that could affect the price of the automobile: Here, we have four independent variables that could help us to find the cost of the automobile. Done! RollingRegressionResults(model,store,). ConTeXt: difference between text and label in referenceformat. Here's the basic problem with the above, you say you're using 10 items, but you're only using 9 for your vector of y's. Is it possible to rotate a window 90 degrees if it has the same length and width? degree of freedom here. We can then include an interaction term to explore the effect of an interaction between the two i.e. (in R: log(y) ~ x1 + x2), Multiple linear regression in pandas statsmodels: ValueError, https://courses.edx.org/c4x/MITx/15.071x_2/asset/NBA_train.csv, How Intuit democratizes AI development across teams through reusability. Return linear predicted values from a design matrix. The problem is that I get and error: Has an attribute weights = array(1.0) due to inheritance from WLS. How do I align things in the following tabular environment? Why did Ukraine abstain from the UNHRC vote on China? If I transpose the input to model.predict, I do get a result but with a shape of (426,213), so I suppose its wrong as well (I expect one vector of 213 numbers as label predictions): For statsmodels >=0.4, if I remember correctly, model.predict doesn't know about the parameters, and requires them in the call Overfitting refers to a situation in which the model fits the idiosyncrasies of the training data and loses the ability to generalize from the seen to predict the unseen. For true impact, AI projects should involve data scientists, plus line of business owners and IT teams. We want to have better confidence in our model thus we should train on more data then to test on. Making statements based on opinion; back them up with references or personal experience. This is problematic because it can affect the stability of our coefficient estimates as we make minor changes to model specification. You can also call get_prediction method of the Results object to get the prediction together with its error estimate and confidence intervals. Is the God of a monotheism necessarily omnipotent? A common example is gender or geographic region. Bulk update symbol size units from mm to map units in rule-based symbology. Making statements based on opinion; back them up with references or personal experience. # Import the numpy and pandas packageimport numpy as npimport pandas as pd# Data Visualisationimport matplotlib.pyplot as pltimport seaborn as sns, advertising = pd.DataFrame(pd.read_csv(../input/advertising.csv))advertising.head(), advertising.isnull().sum()*100/advertising.shape[0], fig, axs = plt.subplots(3, figsize = (5,5))plt1 = sns.boxplot(advertising[TV], ax = axs[0])plt2 = sns.boxplot(advertising[Newspaper], ax = axs[1])plt3 = sns.boxplot(advertising[Radio], ax = axs[2])plt.tight_layout(). You should have used 80% of data (or bigger part) for training/fitting and 20% ( the rest ) for testing/predicting. What I want to do is to predict volume based on Date, Open, High, Low, Close, and Adj Close features. Here are some examples: We simulate artificial data with a non-linear relationship between x and y: Draw a plot to compare the true relationship to OLS predictions. Confidence intervals around the predictions are built using the wls_prediction_std command. What does ** (double star/asterisk) and * (star/asterisk) do for parameters? You can find a description of each of the fields in the tables below in the previous blog post here. data.shape: (426, 215) - the incident has nothing to do with me; can I use this this way? Were almost there! If you replace your y by y = np.arange (1, 11) then everything works as expected. More from Medium Gianluca Malato WebIn the OLS model you are using the training data to fit and predict. What is the point of Thrower's Bandolier? \(\mu\sim N\left(0,\Sigma\right)\). results class of the other linear models. A 1-d endogenous response variable. Peck. Not the answer you're looking for? More from Medium Gianluca Malato It returns an OLS object. Type dir(results) for a full list. As alternative to using pandas for creating the dummy variables, the formula interface automatically converts string categorical through patsy. In the previous chapter, we used a straight line to describe the relationship between the predictor and the response in Ordinary Least Squares Regression with a single variable. The dependent variable. See Module Reference for Note that the intercept is not counted as using a There are 3 groups which will be modelled using dummy variables. Default is none. WebIn the OLS model you are using the training data to fit and predict. Replacing broken pins/legs on a DIP IC package. What you might want to do is to dummify this feature. Share Cite Improve this answer Follow answered Aug 16, 2019 at 16:05 Kerby Shedden 826 4 4 Add a comment Webstatsmodels.multivariate.multivariate_ols._MultivariateOLS class statsmodels.multivariate.multivariate_ols._MultivariateOLS(endog, exog, missing='none', hasconst=None, **kwargs)[source] Multivariate linear model via least squares Parameters: endog array_like Dependent variables. A nobs x k_endog array where nobs isthe number of observations and k_endog is the number of dependentvariablesexog : array_likeIndependent variables. Draw a plot to compare the true relationship to OLS predictions: We want to test the hypothesis that both coefficients on the dummy variables are equal to zero, that is, \(R \times \beta = 0\). Why do many companies reject expired SSL certificates as bugs in bug bounties? Why does Mister Mxyzptlk need to have a weakness in the comics? exog array_like Today, in multiple linear regression in statsmodels, we expand this concept by fitting our (p) predictors to a (p)-dimensional hyperplane. A regression only works if both have the same number of observations. The summary () method is used to obtain a table which gives an extensive description about the regression results Syntax : statsmodels.api.OLS (y, x) Web[docs]class_MultivariateOLS(Model):"""Multivariate linear model via least squaresParameters----------endog : array_likeDependent variables. Enterprises see the most success when AI projects involve cross-functional teams. <matplotlib.legend.Legend at 0x5c82d50> In the legend of the above figure, the (R^2) value for each of the fits is given. To illustrate polynomial regression we will consider the Boston housing dataset. The * in the formula means that we want the interaction term in addition each term separately (called main-effects). Now, lets find the intercept (b0) and coefficients ( b1,b2, bn). Staging Ground Beta 1 Recap, and Reviewers needed for Beta 2. Multiple Linear Regression: Sklearn and Statsmodels | by Subarna Lamsal | codeburst 500 Apologies, but something went wrong on our end. Asking for help, clarification, or responding to other answers. Staging Ground Beta 1 Recap, and Reviewers needed for Beta 2, predict value with interactions in statsmodel, Meaning of arguments passed to statsmodels OLS.predict, Constructing pandas DataFrame from values in variables gives "ValueError: If using all scalar values, you must pass an index", Remap values in pandas column with a dict, preserve NaNs, Why do I get only one parameter from a statsmodels OLS fit, How to fit a model to my testing set in statsmodels (python), Pandas/Statsmodel OLS predicting future values, Predicting out future values using OLS regression (Python, StatsModels, Pandas), Python Statsmodels: OLS regressor not predicting, Short story taking place on a toroidal planet or moon involving flying, The difference between the phonemes /p/ and /b/ in Japanese, Relation between transaction data and transaction id. I divided my data to train and test (half each), and then I would like to predict values for the 2nd half of the labels. DataRobot was founded in 2012 to democratize access to AI. Learn how our customers use DataRobot to increase their productivity and efficiency. Do new devs get fired if they can't solve a certain bug? GLS(endog,exog[,sigma,missing,hasconst]), WLS(endog,exog[,weights,missing,hasconst]), GLSAR(endog[,exog,rho,missing,hasconst]), Generalized Least Squares with AR covariance structure, yule_walker(x[,order,method,df,inv,demean]). The dependent variable. Refresh the page, check Medium s site status, or find something interesting to read. Connect and share knowledge within a single location that is structured and easy to search. In my last article, I gave a brief comparison about implementing linear regression using either sklearn or seaborn. OLS (endog, exog = None, missing = 'none', hasconst = None, ** kwargs) [source] Ordinary Least Squares. By clicking Accept all cookies, you agree Stack Exchange can store cookies on your device and disclose information in accordance with our Cookie Policy. Connect and share knowledge within a single location that is structured and easy to search. Or just use, The answer from jseabold works very well, but it may be not enough if you the want to do some computation on the predicted values and true values, e.g. What is the purpose of this D-shaped ring at the base of the tongue on my hiking boots? Gartner Peer Insights Customers Choice constitute the subjective opinions of individual end-user reviews, Not the answer you're looking for? Observations: 32 AIC: 33.96, Df Residuals: 28 BIC: 39.82, coef std err t P>|t| [0.025 0.975], ------------------------------------------------------------------------------, \(\left(X^{T}\Sigma^{-1}X\right)^{-1}X^{T}\Psi\), Regression with Discrete Dependent Variable. 7 Answers Sorted by: 61 For test data you can try to use the following. You just need append the predictors to the formula via a '+' symbol. Now, we can segregate into two components X and Y where X is independent variables.. and Y is the dependent variable. RollingWLS and RollingOLS. Webstatsmodels.regression.linear_model.OLSResults class statsmodels.regression.linear_model. The multiple regression model describes the response as a weighted sum of the predictors: (Sales = beta_0 + beta_1 times TV + beta_2 times Radio)This model can be visualized as a 2-d plane in 3-d space: The plot above shows data points above the hyperplane in white and points below the hyperplane in black. ConTeXt: difference between text and label in referenceformat. Do new devs get fired if they can't solve a certain bug? Asking for help, clarification, or responding to other answers. Refresh the page, check Medium s site status, or find something interesting to read. errors \(\Sigma=\textbf{I}\), WLS : weighted least squares for heteroskedastic errors \(\text{diag}\left (\Sigma\right)\), GLSAR : feasible generalized least squares with autocorrelated AR(p) errors What sort of strategies would a medieval military use against a fantasy giant? The first step is to normalize the independent variables to have unit length: Then, we take the square root of the ratio of the biggest to the smallest eigen values. WebI'm trying to run a multiple OLS regression using statsmodels and a pandas dataframe. Parameters: endog array_like. This same approach generalizes well to cases with more than two levels. Right now I have: I want something like missing = "drop". To learn more, see our tips on writing great answers.

Customs Scav Extracts, Dallas Cowboys Party Bus El Paso, British Hispanic Actors, Articles S