subgroupsem

This package is used to apply subgroup discovery in combination with structural equation modeling, SEM.

Prerequisites

• R(>= 3.4), R(>4.0.0) if you want to use packrat
• Python3

Ideally one should have a package to create a virtual environments. We recommend having Virtualenv and the R package packrat.

Installing

Install R packages: Optional first step in the directory of your project:

install.packages('packrat')
packrat::init()

Second step:

install.packages('lavaan')
install.packages('plyr')

Install subgroup_sem: Optional first step in the directory of your project:

Virtualenv env
source env/bin/activate

Second step, installing subgroup_sem:

pip install -e .
pip install -e pysubgroup

If it doesn't work because of trouble installing rpy2, try to install that package first

pip install rpy2

How to use

############################################################################################
# load important packages
############################################################################################

import pysubgroup as ps
from subgroup_sem import SEMTarget, TestQF

############################################################################################
# load dataset
############################################################################################

from subgroup_sem.tests.DataSets import get_artificial_data
data = get_artificial_data()

############################################################################################
# define R model and the constraints for the Wald test in lavaan sintax, 
# we use vectors c(x1, x2) to compute the model for the subgroup and the complement 
############################################################################################

model = ('# direct effect \n'
        'Y ~ c(c1,c2)*X \n'
        '# mediator \n'
        'M ~ c(a1,a2)*X \n'
        'Y ~ c(b1,b2)*M \n'
        '# indirect effect (a*b) \n'
        'indirect1 := a1*b1 \n' 
        'indirect2 := a2*b2 \n'
        '# total effect \n'
        'total1 := c1 + (a1*b1) \n'
        'total2 := c2 + (a2*b2) \n'
        '# direct effect \n'
        'direct1 := c1 \n'
        'direct2 := c2 \n'
        '# rest \n'
        'Y ~~ c(r1_1,r1_2)*Y \n'
        'X ~~ c(r2_1,r2_2)*X \n'
        'M ~~ c(r3_1,r3_2)*M \n'
        'Y ~ c(r4_1,r4_2)*1 \n'
        'X ~ c(r5_1,r5_2)*1 \n'
        'M ~ c(r6_1,r6_2)*1')

wald_test_contstraints = 'a1==a2 \n b1==b2 \n c1==c2'

############################################################################################
# define quality function
############################################################################################

def q(WT_score):
    return WT_score

############################################################################################
# define and perform subgroup discovery task
############################################################################################

target = SEMTarget (data, model, wald_test_contstraints)
searchSpace = ps.create_selectors(data, ignore=["X", "Y", "M"])
task = ps.SubgroupDiscoveryTask(data, target, searchSpace, result_set_size=10, depth=2, 
                                qf=TestQF(quality_function = q, parameters_list = ['WT_score'], parameters_type_dic={}))
result = ps.SimpleDFS().execute(task)

############################################################################################
# print the results
############################################################################################

for (q, sg) in result.to_descriptions():
    print(str(q) + ":\t" + str(sg))

The first lines import the necessary packages and data. After we define the model and the constraints for the Wald test in lavaan sintax. Next we define our quality function. This data is used to define the target and the task. The class TestQF takes as arguments the quality function, next are the arguments in order that will be passed to this function. The elements of the list can be chosen from:

• 'size_sg' for subgroup size
• parameters listed in fitMeasures(fit)
• parameters listed in parameterEstimates(fit)
• Wald Test scores: 'WT_score', 'WT_p', 'WT_df', 'WT_se'

parameters_type_dic: For all the parameter in parameterEstimates(fit) one has the option of choosing the type ('est' for the value itself, 'z' for the z score, 'pvalue' for the p value). For example:

parameters_type_dic = {direct1:'est', total1:'pvalue'}

If parameters_type_dic remains empty the value of the parameter will be taken. For more information about the options visit the lavaan website.

Authors

• Florian Lemmerich - Initial work - Pysubgroup
• Christoph Kiefer
• Benedikt Langenberg
• Jeffry Cacho Aboukhalil
• Axel Mayer
• Felix Stamm

License

We are happy about anyone using this software. Thus, this work is put under an Apache license. However, if this constitutes any hindrance to your application, please feel free to contact us, we are sure that we can work something out.

Copyright 2016-2020 Florian Lemmerich, Christoph Kiefer, Benedikt Langenberg, Jeffry Cacho Aboukhalil, Axel Mayer, Felix Stamm

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Cite

...