DataFITR: An Open, Guided Input Modeling Tool for Creating

Simulation-Based Digital Twins

Lekshmi P.

, Tushar Lone

and Neha Karanjkar

Indian Institute of Technology Goa, India

Keywords:

Input Modeling, Simulation, Digital Twins, Open Source, Graphical User Interface.

Abstract:

Input Modeling (IM) is a critical step in the process of building simulation-based digital twins. It involves

selecting a family of distributions to model the observed data and ﬁnding the distribution parameter values that

best ﬁt the data. Subsequently, random variates adhering to the selected distribution can be generated to create

a simulation-based digital twin of the system. For complex systems, IM can be a nuanced process involving

a series of decisions that require visual feedback at each step. There is currently a dearth of open, GUI-based

tools for aiding the non-expert user in the process of IM. This paper presents DataFITR, a GUI-based, open

Input Modeling tool we have developed for guiding the non-expert user through the steps of input modeling

and automating several intermediate tasks. DataFITR is cloud-hosted with a web-based user interface. The

user can upload data as a ﬁle and the tool guides the user through the IM process by suggesting types and

suitable distributions for each observed variable. It generates multiple goodness-of-ﬁt measures for a large set

of standard discrete and continuous distributions and can also support arbitrary (non-standard) distributions

using a Kernel Density Estimation approach. DataFITR also assists in exploratory data analysis by providing

various statistical properties of the observed data and in ﬁnding correlations between output measures. Once

a matching distribution is found, the tool generates Python code for producing random variates from the

matching distribution, which can be directly inserted into a simulation model. In this paper, we describe the

DataFITR tool and its features, and compare it with existing open libraries and tools for assisting IM. We

present a simulation case study of a bottling plant to demonstrate the utility of the DataFITR tool in building

simulation-based digital twins.

1 INTRODUCTION

A Digital Twin refers to a virtual replica or model

of a physical system whose state is kept in sync with

the real system via a continuous stream of observa-

tions or data from the real system. A digital twin may

also be used to control the real system via feedback

paths. Figure 1 illustrates the main components (such

as sensing and actuation, data aggregation, simula-

tion and dashboarding) that make up a digital twin.

The use of digital twins has the potential to revo-

lutionize sectors such as manufacturing, healthcare,

urban planning, and transportation by enabling real-

time decision-making using Internet of Things (IoT)

technology and real-time analytics (Kritzinger et al.,

2018).

A digital twin can either be purely data-driven or

https://orcid.org/0000-0001-5464-6032

https://orcid.org/0000-0003-0008-0429

https://orcid.org/0000-0003-3111-1435

simulation-based. In a purely data-driven digital twin,

the behavior of the system and future predictions are

derived solely from data. In contrast, a simulation-

based digital twin can be used when the modeler has

prior knowledge or can make reasonable assumptions

about the behavior of the system. The system behav-

ior, structural components and/or state transitions in

the system are described by the modeler to create a

parameterized simulation model and the exact values

of these parameters are derived from observed data.

Depending on the type of real system under consid-

eration, simulation models can either be deterministic

or stochastic, and are typically implemented using ei-

ther a discrete-event simulation approach or continu-

ous simulation paradigms.

Input Modeling (IM) is a critical step in the cre-

ation of digital twins. When performed incorrectly or

without due care, the resulting model is prone to suf-

fer from the Garbage-In-Garbage-Out problem and

may yield incorrect insights or predictions. For a dig-

P., L., Lone, T. and Karanjkar, N.

DataFITR: An Open, Guided Input Modeling Tool for Creating Simulation-Based Digital Twins.

DOI: 10.5220/0012082600003546

In Proceedings of the 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2023), pages 279-286

ISBN: 978-989-758-668-2; ISSN: 2184-2841

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

279

ital twin the process of IM needs to be automated

as the model needs to remain continuously in sync

with incoming data. Traditionally, IM involves deriv-

ing simulation parameters (which can either be con-

stants or parameters of an appropriate probability dis-

tribution) from data corresponding to multiple mea-

sured entities. For example, in a manufacturing plant,

one might have a continuous stream of data corre-

sponding to attributes such as time taken by a ma-

chine to perform certain kinds of tasks, the periodi-

cally monitored levels of inventory or raw material,

energy consumption of a machine per-task etc. If a

quantity (such as the time taken per-task) is assumed

to be random, it can be modeled by generating ran-

dom variates inside a stochastic simulation model,

where the parent distribution of this random vari-

able can be obtained from observed data. IM refers

to the systematic process of selecting the appropri-

ate distributions for each physical quantity and ﬁtting

their parameters to best match the observed data. For

time-independent data, this process involves select-

ing the right distribution and ﬁnding the maximum

likelihood estimators for the distribution using data.

Several Goodness of ﬁt measures can be used to de-

scribe the extent of match between the selected distri-

bution and the observed data. These steps need to be

repeated until a reasonable ﬁt is achieved. For time-

dependent data, this involves selecting the appropri-

ate models (stationary/non-stationary) and mimicking

the time-dependence using mathematical models such

as Markov Chains or Moving-Average (MA) models

(Banks et al., 2010). In a real system, it is often the

case that some of the measured quantities are corre-

lated. For example, the energy consumption of a ma-

chine for performing a certain task and the time taken

to perform that task may be positively correlated. This

correlation needs to be identiﬁed right at the outset,

and modeled using multivariate distributions for gen-

erating random variates in the simulation model.

Thus the task of IM is nuanced and may require

time and effort. However, with increasing ubiquity of

digital twins and their creation or use by non-expert

users, it becomes necessary to have tools that as-

sist and guide the user in IM or automate some as-

pects of this task. Recently, deep-learning based ap-

proaches such as generative neural architectures have

been proposed to automate IM (Cen et al., 2020).

Such methods may require a large amount of train-

ing data. For cases where a modeler has some prior

knowledge about the system behavior, traditional ap-

proaches may be better suited and efﬁcient as stan-

dard distributions often end up mimicking the ob-

served data really well. While there exist commer-

cial tools such as ExpertFit (Law, 2020) and Stat::Fit

Figure 1: Components of a digital twin.

(Software, 2022) and libraries in programming lan-

guages (summarized in the next section) which can

assist the user in IM by ﬁtting distributions to data,

there is currently a dearth of free/open IM tools tar-

geted for the non-expert user. Desirable properties for

an IM tool include a GUI interface, assistive features,

support for a variety of goodness of ﬁt tests, and the

ability to ﬁt correlated data and generate random vari-

ates. While some commercial tools provide some of

these features, there is a lack of open/free tools for

IM that offer a GUI-front-end and are targeted for the

non-expert user.

This paper describes DataFITR - an open, GUI-

based, cloud-hosted tool we have developed for as-

sisting the non-expert user in input modeling. The

tool is freely accessible via a browser at the URL:

https://dataﬁtr.streamlit.app. The user can upload

their data as a csv (comma-separated-value) ﬁle. The

tool then guides the user through the input model-

ing process in a step-wise manner while providing vi-

sual aids (such as data plots, histograms and color-

coded correlation matrices) at each step. It automati-

cally generates plots of the marginal distributions and

a statistical summary of the data. The user can do ex-

ploratory data analysis of the columns in the data set

and then proceed to ﬁnd the distributions that best ﬁt

the data. The tool supports a large set of standard con-

tinuous and discrete distributions. It also implements

and reports multiple goodness of ﬁt measures such as

Kolmogorov-Smirnov (KS) test metric, Chi-squared

(χ

) test measure, and Sum of Squares Error (SSE).

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

280

For data that does not match any standard distribution,

the DataFITR tool can generate arbitrary uni-variate

distributions to match the histogram using the Kernel

Density Estimation (KDE) approach (Parzen, 1962).

Most importantly, once a desired distribution is found,

the tool automatically produces Python code for ran-

dom variate generation corresponding to the selected

distribution. This code can be directly copied into a

stochastic simulation model to generate the random

variates. DataFITR has been written in Python, and

a GUI front-end is created using the Streamlit library

(Streamlit, 2022). The tool is currently cloud-hosted

on the Streamlit community cloud. DataFITR cur-

rently supports time-independent models (with a large

set of standard distributions or arbitrary distribution),

and for correlated data multi-variate Gaussian distri-

butions are currently supported. We plan to add sup-

port for time-dependent models and arbitrary multi-

variate distributions in future versions.

The rest of this paper is organised as follows: In

Section 2 we provide a brief overview of open li-

braries and existing tools available for building digital

twins with a focus on IM. In Section 3, we describe

the features, usage ﬂow and details of the DataFITR

tool. In Section 4 we present a simulation case study

of a bottling plant which serves to highlight the util-

ity of the tool. In this case study, we generate data

using a known reference model (a discrete-event sim-

ulation model) of a bottling plant, and use this data

and some knowledge about the real system to create

a matched model automatically using the DataFITR

tool. We present results showing the extent of match

between the original reference model and the matched

model in terms of the system parameters and out-

put/performance measures. Finally, we present con-

clusions and future plans in the last section.

2 RELATED WORK

A broad survey of tools and processes in creating dig-

ital twins is presented in (Fuller et al., 2020). In-

put Modeling (IM) is a critical step and historically

IM techniques have focused on ofﬂine system mod-

els (Cheng, 2017). An overview of IM techniques

for various problem domains is described in (Nelson

and Yamnitsky, 1998). Commercial software tools

such as ExpertFit (Law, 2020) and Stat::Fit (Soft-

ware, 2022) support in IM by identifying probabil-

ity distributions to ﬁt observed data. These tools also

assist the user in selecting a distribution when data

is unavailable based on system knowledge, for as-

pects such as task times and equipment failures. XL-

STAT (Lumivero, 2022) is a commercial Excel-based

tool that can be used for IM. Aside from commer-

cial tools, a few libraries in popular programming lan-

guages such as Python and R exist for ﬁtting proba-

bility distributions to data. Distﬁt (Taskesen, 2020)

and ﬁtter (Cokelaer, 2020) are two examples of open

Python-based libraries. Both can be used to ﬁt stan-

dard uni-variate distributions. ﬁtteR (Boenn, 2022) is

an R-based version for ﬁtting distributions to empir-

ical data. Distribution ﬁtter (Distributionﬁtter, 2022)

is a python based GUI application which is built using

the ﬁtter package. It can be used to ﬁt univariate dis-

tributions. Distribution Analyser (DistributionAnal-

yser, 2022) is another Python-based application that

helps users analyze univariate distributions. It also

allows the users to ﬁt the data into univariate distri-

butions. While these are libraries that can be used via

interface routines, the DataFITR tool described in this

paper is a GUI-based tool that does not require any

programming for its use. Table 1 summarizes the dif-

ferent features and scope of these libraries along with

the DataFITR tool proposed in this paper.

3 DataFITR: FEATURES AND

USAGE

DataFITR is open-source (released as a public repos-

itory on GitHub at (Lekshmi P, 2023)). It is currently

hosted on Streamlit public cloud and freely accessi-

ble via a browser at https://dataﬁtr.streamlit.app. The

user can upload the data in a csv (comma separated

value) format where each column corresponds to a

single measured quantity and the ﬁrst row is assumed

to contain the names of each quantity. For categori-

cal type of data, it is assumed that the data is integer-

valued.

DataFITR currently supports modeling time-

independent, Independent and Identically Distributed

(IID) data where each measured quantity (column in

the data ﬁle) is independent. It also supports mod-

eling the case where some of the columns are corre-

lated. For the multivariate case the tool currently sup-

ports only Gaussian distributions, and the ability to

ﬁt arbitrary multivariate distributions is planned to be

implemented in future. The side panel in the tool can

be used for selecting one of these cases for the IM

ﬂow. To ﬁt time-independent IID data, user can up-

load the ﬁle on the on the page corresponding to the

time-independent data. Once the user uploads a ﬁle,

the tool automatically identiﬁes the column headings

and whether each column corresponds to real-valued

(continuous) or integer-valued (categorical) data, cre-

ates histograms showing the marginal distributions of

each columns. It also generates a statistical summary

DataFITR: An Open, Guided Input Modeling Tool for Creating Simulation-Based Digital Twins

281

Table 1: Comparison of tools and libraries for IM.

Features Distﬁt ﬁtter Distribution Analyser Distribution ﬁtter DataFITR

GUI , Visualization × × ✓ ✓ ✓

Random Variable Gen × × ✓ × ✓

Arbitrary distributions × × × × ✓

Support for Multivariate × × × × ✓

Gaussian distribution

Num Goodness of ﬁt 3 1 1 1 3

measures used

Std distributions 89 80 97 80 100

Dependencies Scipy Scipy Scipy, Streamlit Scipy, Streamlit Scipy, Streamlit

of each column. The tool then produces a correla-

tion matrix between every pair of columns and points

out the columns with high correlation. This indicates

to the user whether some quantities need to be mod-

eled using the multivariate case. The user can then

proceed with one column at a time to ﬁnd the best-

ﬁtting distributions by conﬁrming the details entered

by the tool based on the data (as shown in Figure 2)

and obtain Python code for generating random vari-

ates from that distribution. The user can also edit the

options entered by the tool. The tool then generates

histograms and plots for visualizing the extent of ﬁt

for a selected set of standard distributions ( as shown

in Figure 3) and also generates a distribution match-

ing the histogram using the KDE approach for data

that does not resemble any standard distribution.

DataFITR supports 97 standard continuous dis-

tributions and 3 discrete distributions (Poisson, Bi-

nomial and Geometric ). For continuous distributions,

users can either select the set of popular/common

continuous distributions (such as Normal, Triangu-

lar, Uniform, Exponential, Lognormal, Weibull, and

Gamma) or all of the 97 standard continuous distri-

butions for performing a best ﬁt with the data. For

the selected set of standard distributions, the tool ﬁrst

ﬁnds the distribution parameters (maximum likeli-

hood estimators) to best ﬁt the data, and then for each

distribution it reports three goodness of ﬁt measures

in a table. The Kolmogorov–Smirnov (KS) statistic

quantiﬁes the distance between the empirical distribu-

tion function of the sample and the cumulative distri-

bution function of the reference distribution. The Chi-

squared (χ

) test measures the distance between the

normalized data histogram and the probability den-

sity function of the reference distribution. The Sum of

Squared Errors (SSE) metric is the sum of squares of

the differences between the probability density func-

tion (pdf) of the reference distribution and the data

histogram, with a speciﬁc bin size selected for the

data. Once a matching distribution is found, the user

can select the desired distribution and the tool gen-

erates Python code that can be directly copied to the

simulation model as shown in Figure 4. Further, the

tool also has a feature for generating Python code for

random variate generation for selected standard distri-

butions. For this the user has to supply the distribution

name and the parameters. Finally, the summary of the

matched distributions can be seen by the user in the

View Output Summary tab.

4 CASE STUDY

Table 2: Model parameters and their values used in the ref-

erence model and estimated values of the model parameters

in the matched model.

Parameter

(unit)

Reference

model

(distribution,

parameter)

Matched

model

(distribution,

parameter)

per job delay of

cap unit (mins)

uniform

a=1.25

b=2.00

uniform

a=1.250

b=1.999

per job delay of

bottle unit(mins)

exponential

λ=0.50

exponential

λ=0.500

per job delay of ﬁlling

capping unit(mins)

normal

µ=2.75

σ=0.05

normal

µ=2.7500

σ=0.0505

interval of reﬁll

raw materials(mins)

uniform

a=60

b=90

uniform

a=60.52

b=88.99

amount of plastic

for cap unit(numbers)

constant

amount of plastic

for bottle unit(numbers)

constant

amount of drink

for ﬁlling (millilitres)

constant

200

constant

200

We present a case study that illustrates the use of the

DataFITR tool for creating simulation-based digital

twins. In this case study we assume that the real sys-

tem is represented by a reference model of a bottling

plant containing various processes and components

as illustrated in Figure 5. We have built a detailed

Discrete-Event simulation model of this system using

the SimPy library (Klaus G. M

uller, 2020) in Python.

Simulating this reference model generates data mim-

icking the data stream that would have been generated

by a manufacturing IoT infrastructure in a real man-

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

282

Figure 2: DataFITR: Choosing the datatype, the number of bins and the set of distributions to ﬁt.

Figure 3: DataFITR: Results of IM for the selected column.

ufacturing system. This data is automatically saved

into a csv ﬁle and serves as an input to the DataFITR

tool. We then use the generated data along with the

DataFITR tool to build another model (a matched

model) of the system which would serve as the under-

lying model in a simulation-based digital twin. Be-

cause the reference model is known, we can then re-

port the extent of match between the reference model

and the matched model, serving as a validation exer-

cise and an illustration of the utility for the DataFITR

DataFITR: An Open, Guided Input Modeling Tool for Creating Simulation-Based Digital Twins

283

Figure 4: DataFITR: Random variate generation for the ﬁt distribution.

Table 3: Comparison of output performance measures aver-

aged across 750 simulation runs.

Parameter Reference Matched Error

(unit) model model in %

Throughput 18.92000 18.91000 0.035

(Num products/hour)

Average cycletime 6.48700 6.48800 0.006

(mins)

Resource utilisation 0.86970 0.86950 0.004

of cap unit(%)

Resource utilisation 0.86863 0.86861 0.002

of Bottle unit(%)

Resource utilisation 0.86720 0.86710 0.005

of ﬁlling-capping

unit (%)

tool.

Figure 5 shows a bottling plant with four stages

connected in an assembly line fashion. The four

stages are cap-making, bottle-making, bottle-ﬁlling,

capping, and reﬁlling. The green blocks supply the

raw materials and are real-valued components and the

red blocks are the integer-valued components. The

yellow blocks are the machines where the active man-

ufacturing process takes place. The model’s parame-

ters are the per-job delays at each unit, the amount of

plastic required to make a cap, a bottle and the amount

of drink required to ﬁll a bottle. These are shown in

blue blocks. The raw materials are stored separately

and their levels are monitored continuously. A re-

ﬁll process replenishes the raw materials periodically.

There is downtime for all machines while the reﬁlling

process takes place. The caps and bottle bodies are

manufactured and stored separately. The drink mix

is ﬁlled into bottles and capped at the next stage. Fi-

nally, these bottles are moved to a store and can be

supplied to various distribution units.

The process delays are modeled as random vari-

ables with parameters summarized in Table 2. The

reference model generates timestamped data which is

similar to what an IoT framework in a manufacturing

unit would have generated. This data becomes the in-

put to the DataFITR and the generated code from the

DataFITR is used for generating random variables in

the matched model. After this, both the simulation

models are run and a comparison of the output per-

formances is presented to evaluate the extent of the

match of both models.

4.1 Model Parameters and

Implementation

The simulation model parameters identiﬁed from the

reference model are listed in Table 2. The ﬁrst four

parameters are variable and the remaining three are

constant. This reference model is implemented in a

python package called SimPy. It is a framework for

developing Discrete Event Simulation. It is based on

processes which are simple python generator func-

tions. In the reference model, separate processes are

developed to simulate a cap-making unit, a bottle-

making unit, a ﬁlling and capping unit, and a reﬁlling

unit. Units that are used to store the raw materials are

modeled as containers which is a shared resource. In

the reference model, plastic and drink mix are stored

in a shared resource called a container with a capacity

of 1000 no.s and 25 L, respectively. The manufac-

tured caps, bottle bodies, and the ﬁlled and capped

bottles are stored in another type of resource called

a Store. In one simulation, the plant is run for 1000

hours. 750 such simulations are run, and the parame-

ters listed in Table 3 are estimated from the simulation

runs.

4.2 Performance Estimation and

Results

The output performance measures (such as system

throughput, average cycle time and resource utiliza-

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

284

Figure 5: Schematic of a bottling plant.

Figure 6: Plot showing the extent of match between the reference model and matched model. Plot on the left shows the

variation of absolute value of throughput with respect to the number of simulation runs and the plot on the right shows the

variation of the relative standard error in throughput with respect to the number of simulation runs.

tion of each of the processing machines) are estimated

after averaging the results from multiple stochastic

simulation runs. The number of simulation runs re-

quired to get an estimate with reasonable accuracy is

determined after calculating the relative standard er-

ror. Figure 6 shows the extent of match between the

reference model and matched model. The x-axis is

the number of simulation runs, and y-axis are abso-

lute value of throughput and relative standard error in

throughput. The result from the reference model are

compared with the results from the matched model

and summarized in tables 2 and 3. We observe that

the percentage error between the two models is small.

In this exercise, we assumed that a modeler has prior

knowledge about the structure and components in the

model and the DataFITR tool was simply used to

match the parameter values of one model to another.

Inferring the model structure/behavior itself from data

DataFITR: An Open, Guided Input Modeling Tool for Creating Simulation-Based Digital Twins

285

is an interesting problem relevant to rapid deployment

of digital twins. The case study highlights the ease of

use of the DataFITR tool.

5 CONCLUSIONS

This paper presented DataFITR, an open, cloud-based

tool for assisted input modeling. The tool has features

to perform automatic exploratory data analysis, ﬁnd

correlations between measured quantities, and iden-

tify distributions and their parameters to best match

the observed data using multiple goodness of ﬁt mea-

sures. At each step the tool generates visual aids (such

as histograms and plots) to help the user select an ap-

propriate model and the tool also takes input from the

user such as prior knowledge about the possible set

of distributions likely to mimic observed data. For

data that cannot be modeled using standard distri-

butions, the tool supports generating arbitrary den-

sity functions using the KDE approach. Most im-

portantly, DataFITR automatically generates Python

code for producing random variates of the matching

distribution which the user may directly copy into a

simulation model of the system. We have also pre-

sented a comparison of our tool with other open-

source packages. A case study of a bottling plant

was presented to illustrate the utility of the tool. Cur-

rently, the tool implements models for iid data. Fit-

ting arbitrary multi-variate distributions and modeling

time-dependent data are features planned to be imple-

mented in future versions.

REFERENCES

Banks, J., II, J. S. C., Nelson, B. L., and Nicol, D. M.

(2010). Discrete-Event System Simulation, 5th New

Internatinal Edition. Pearson Education.

Boenn, M. (2022). ﬁtteR: Systematic ﬁt of hundreds

of theoretical univariate distributions to empirical

data. https://cran.r-project.org/web/packages/ﬁtteR/

index.html.

Cen, W., Herbert, E. A., and Haas, P. J. (2020). Nim: Mod-

eling and generation of simulation inputs via genera-

tive neural networks. In 2020 Winter Simulation Con-

ference (WSC), pages 584–595.

Cheng, R. (2017). History of input modeling. In 2017 Win-

ter Simulation Conference (WSC), pages 181–201.

Cokelaer, T. (2020). ﬁtter: A Python library for ﬁtting prob-

ability distributions to data. https://ﬁtter.readthedocs.

io/en/latest/index.html.

DistributionAnalyser (2022). Distribution Anal-

yser:An app to interactively explore and ﬁt

continuous distribution functions. https:

//rdzudzar-distributionanalyser-main-45cc69.

streamlit.app/.

Distributionﬁtter (2022). Distribution ﬁtter: An app to

compare multiple distributions and ﬁnd the best one

that ﬁts your data. https://github.com/rahul-raoniar/

distribution

ﬁtter streamlit app.

Fuller, A., Fan, Z., Day, C., and Barlow, C. (2020). Digi-

tal twin: Enabling technologies, challenges and open

research. IEEE Access, 8:108952–108971.

Klaus G. M

uller, T. V. (2020). SimPy:Discrete-event

simulation framework for Python. https://simpy.

readthedocs.io/en/latest/.

Kritzinger, W., Karner, M., Traar, G., Henjes, J., and Sihn,

W. (2018). Digital twin in manufacturing: A cat-

egorical literature review and classiﬁcation. IFAC-

PapersOnLine, 51(11):1016–1022. 16th IFAC Sym-

posium on Information Control Problems in Manufac-

turing INCOM 2018.

Law, A. M. (2020). ExpertFIT. http://www.averill-law.com/

distribution-ﬁtting/.

Lekshmi P, Neha Karanjkar, T. L. (2023). DataFITR:

public repository. https://github.com/NehaKaranjkar/

DataFITR.

Lumivero (2022). XLSTAT: DISTRIBUTION FITTING.

https://www.xlstat.com/en/.

Nelson, B. and Yamnitsky, M. (1998). Input modeling tools

for complex problems. In 1998 Winter Simulation

Conference. Proceedings (Cat. No.98CH36274), vol-

ume 1, pages 105–112 vol.1.

Parzen, E. (1962). On estimation of a probability den-

sity function and mode. The Annals of Mathematical

Statistics, 33(3):1065–1076.

Software, G. M. (2022). Stat::Fit Distribution Fitting Soft-

ware. https://www.geerms.com.

Streamlit (2022). Streamlit:A open-source app framework

for Machine Learning and Data Science. https://docs.

streamlit.io/.

Taskesen, E. (2020). Distﬁt: A python package for proba-

bility density ﬁtting of univariate distributions. https:

//erdogant.github.io/distﬁt/pages/html/index.html.

SIMULTECH 2023 - 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

286