Cara menggunakan reverse dummy variables python

Create a set of dummy variables representing the values of one or more variables and interaction terms

This procedure creates a set of 0-1 dummy variables representing the distinct values of one or more variables. It can also create dummies for two- and three-way interaction terms. It is useful, for example, in converting categorical variables into a set of variables appropriate for use in the Regression procedure. It can optionally create macro variables representing sets of dummies.

Requirements

• IBM SPSS Statistics 18 or later and the corresponding IBM SPSS Statistics-Integration Plug-in for Python.

Note: For users with IBM SPSS Statistics version 19 or higher, the SPSSINC CREATE DUMMIES extension is installed as part of IBM SPSS Statistics-Essentials for Python.

Installation intructions

1. Open IBM SPSS Statistics
2. Navigate to Utilities -> Extension Bundles -> Download and Install Extension Bundles
3. Search for the name of the extension and click Ok. Your extension will be available.

Tutorial

Create a set of dummy variables for values of one or more variables

SPSSINC CREATE DUMMIES VARIABLES=variables ROOTNAME1=rootnames ROOTNAME2=rootname ROOTNAME3=rootname

/OPTIONS MAXVARS=n ORDER=A^** or D
MACRONAME1="name" MACRONAME2="name" MACRONAME3="name"
OMITFIRST=YES^** or NO
USEVALUELABELS=NO^** or YES
USEMLS=YES^** or NO

/HELP

^* Required
^** Default

SPSSINC CREATE DUMMIES /HELP displays this help and does nothing else.

Example:

SPSSINC CREATE DUMMIES VARIABLES=job gender
ROOTNAME1 = jobroot genderroot  ROOTNAME2=two
MACRONAME1 = "!job" "!gender".

This command creates a set of 0-1 dummy variables for values of one or more variables. This can include dummies for individual variables, two-way interactions, and three-way interactions. This is useful, for example, when running the Regression procedure and you have categorical variables that needs to be treated as factors. Scale measurement variables can be included and are treated appropriately for interaction terms.

Any cases excluded by the current filter are not considered in determining the set of values.

VARIABLES specifies the variables for which dummies are created. ROOT1 specifies root names for the one-way dummy variables. Specify one root name for each variable listed in VARIABLES. ROOT2 and ROOT3 each specify a single root name for all two-way or three-way effects, respectively. Only orders for which there is a root name are generated.

For one-way dummies, the dummy names will have the form rootname_n. Names for two-way and three-way dummies have the form rootname_m_n, and rootname_m_n_o. Existing variables will be replaced.

For a scale variable, the one-way generated variable is just a copy rather than a set of individual dummies. For two-way dummies, for example, if x is scale and y is categorical with values 1, 2, and 3, the two-way interaction would generate x*(y eq 1), x * (y eq 2), and x*(y=3).

OPTIONS

Specify USEMLS=NO to force all variables to be treated categorically.

Specify MAXVARS=n to limit the number of dummy variables generated to no more than n. This applies only to the one-way interactions.

By default the dummies correspond to ascending values of the input variable. Specify ORDER=D for the reverse order. For interaction dummies, the values are sorted in the order in which the variables are listed in the VARIABLES keyword.

The variable labels for the new variables have forms like varname = value or (varname = value) * (varname = value) or varname * (varname = value).

USEVALUELABELS=YES causes the value label corresponding to the dummy to be used in the label. If a value does not have a label, the value itself is used. USEVALUELABELS=NO causes the values to be used instead.

MACRONAME1=name creates a set of macros whose text is the list of dummy variables. The number of names must match the number of variables. MACRONAME2 and MACRONAME3 specify a single name each for the two- and three-way macros. Note that the macro names must be enclosed in quotes. (Otherwise the system would attempt to expand them.) However, you can use one set of quotes surrounding the entire list as a shortcut.

By default, the first dummy (after apply the order specification) is omitted from the list for each macro . This is convenient in Regression if a constant term is included. Specify OMITFIRST=NO to include all the variables in the macro text.

ROOTNAME is a synonym for ROOTNAME1; MACRONAME is a synonym for MACRONAME1, and VARIABLE is a synonym for VARIABLES.

Pandas is a special tool that allows us to perform complex manipulations of data effectively and efficiently. Inside pandas, we mostly deal with a dataset in the form of DataFrame. DataFrames are 2-dimensional data structures in pandas. DataFrames consist of rows, columns, and data.

Reversing a get dummies encoding

Dummy columns in pandas contain categorical data into dummy or indicator variables. These are used for data analysis. In most cases, this is a feature of any action being described.

To get a dummy column, we must use pandas.get_dummies() method, this method returns all the dummy values of each column passed as a series inside it.

Here, we are given a DataFrame with multiple columns, out of these columns only one column has some variability and other columns only have 0 or 1 value.

Let us say column X has some values and we need to find out for each value of X which column has 1 in front of it.

For this purpose, we will set the index of the DataFrame with the values of column X and use the stack() method with a condition that only if the values are 1. After this, we will reset the index and we will drop all the zeroes from the columns.