=Paper=
{{Paper
|id=Vol-3899/paper9
|storemode=property
|title=Method for analysis and formation of representative text datasets
|pdfUrl=https://ceur-ws.org/Vol-3899/paper9.pdf
|volume=Vol-3899
|authors=Olena Sobko,Olexander Mazurets,Maryna Molchanova,Iurii Krak,Olexander Barmak
|dblpUrl=https://dblp.org/rec/conf/advait/SobkoMMKB24
}}
==Method for analysis and formation of representative text datasets==
https://ceur-ws.org/Vol-3899/paper9.pdf
Method for analysis and formation of representative text
datasets⋆
Olena Sobko1,∗,†, Olexander Mazurets1,†, Maryna Molchanova1,†, Iurii Krak2,3,† and
Olexander Barmak1,†
1 Khmelnytskyi National University, Khmelnytskyi, 11, Institutes str., 29016, Ukraine
2 Taras Shevchenko National University of Kyiv, Kyiv, 64/13, Volodymyrska str., 01601, Ukraine
3 Glushkov Cybernetics Institute, Kyiv, 40, Glushkov ave., 03187, Ukraine
Abstract
The paper is devoted to the creation and approbation of method for analysis and formation of
representative text datasets according to FATE fairness principle for subject areas. The method performs
an analysis of dataset representativeness according to ethical aspects, as result of which a representative
adjustment of the dataset according to ethical aspects is performed. When adjusting the dataset,
optimization problem is solved both for the selection of redundant elements for removal, and for the
formation of requirements for ethical aspects of belonging to each element for data augmentation. To
investigate the effectiveness of the method, software was created that uses machine learning models to
classify texts according to various ethical aspects – age, gender, religion, ethnicity, etc. The obtained
deviations of the sample distributions by ethical aspects classes of dataset, transformed according to the
created method, compared to the ideal representative distribution were: minimum 0.00%, maximum 0.04%,
average 0.02%. The obtained results contribute to improvement of representativeness of text datasets and
fair and unbiased representation of demographic groups in them, which increases trust in decisions made
by artificial intelligence.
Keywords
NLP, data ethical correctness, ethical principles, non-discrimination, text datasets representative 1
1. Introduction
In today's world, numerous solutions using artificial intelligence are being actively developed to
solve various tasks that people face every day. Accordingly, the results generated by artificial
intelligence depend on the training datasets on which they were trained, in other words, the
content of these datasets directly affects the final result. Lack of transparency regarding the
sources and characteristics of the data used to train AI algorithms reduces confidence in the results
obtained. In this case, users are often unable to appreciate the potential biases or discriminatory
elements built into these algorithms. Insufficient awareness of the content of educational datasets
increases the risk of spreading unfair or inaccurate decisions, which can have serious consequences
for individuals and society as a whole [1].
Means for evaluating the representativeness of a textual data set in accordance with the
principles of ethical non-discrimination are currently lacking. This is especially relevant for
AdvAIT-2024: 1st International Workshop on Advanced Applied Information Technologies, December 5, 2024, Khmelnytskyi,
Ukraine - Zilina, Slovakia
∗ Corresponding author.
† These authors contributed equally.
olenasobko.ua@gmail.com(O. Sobko); exe.chong@gmail.com (O. Mazurets);
m.o.molchanova@gmail.com (M. Molchanova); yuri.krak@gmail.com(I. Krak); barmak@gmail.com (O. Barmak)
0000-0001-5371-5788 (O. Sobko); 0000-0002-8900-0650 (O. Mazurets);
0000-0001-9810-936X (M. Molchanova); 0000-0002-8043-0785 (I. Krak); 0000-0003-0739-9678 (O. Barmak)
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
socially important and sensitive tasks according to SDG3 (good health and well-being), SDG4
(quality education) and SDG16 (peace, justice, and strong institutions), for example, detecting
cyberbullying, determining the emotional state of people based on text messages, etc. The lack of
attention to ethical components when creating and using datasets leads to bias in algorithms,
which negatively affects the fairness and reliability of the decisions made [2].
Well-known datasets for training neural networks, for example [3, 4], are actively used by
researchers, because they have a large amount of data, but they were not validated by the authors
regarding representativeness according to the principle of fairness, and therefore, the use of such
datasets for training artificial intelligence algorithms may potentially violate ethical principles and,
hence, have a low reliability of the decisions made.
The representativeness of the data in the datasets not only affects the accuracy of the results
and models, but is also closely related to the principles of FATE (Fairness, Accountability,
Transparency, Ethics) in the use of data and development of artificial intelligence technologies. If
dataset does not include adequate representation of all social, demographic, or cultural groups, it
can lead to discriminatory patterns that prioritize one group over another, so are not fair. The
representativeness of datasets according to ethical principle of FATE is achieved by correct
balancing according to various ethical aspects: gender, religious, age, etc. [5].
The main contribution of the paper is the development and validation of an approach to the
analysis and formation of representative text samples of data according to the principle of fairness
of FATE for subject areas.
Further, in chapter 2, a review of works related to the topic of the study, namely the formation
of representative text samples and the issue of impartial representation of demographic groups
according to the principle of justice, is carried out. Chapter 3 offers a description of the method of
analysis and formation of representative samples of text data, the datasets used for further
experimental studies of the effectiveness of the given method are given and described. Chapter 4
contains an experimental study. Section 5 presents the results and discussion. Chapter 6 concludes
the work.
2. Related work
Many works have been devoted to the study of the representativeness of text samples and the fair
and unbiased representation of demographic groups in them, since the concepts of
representativeness, fairness and impartiality are important in the creation of ethical and fair
machine learning models [6]. Natural language processing tools are widely used for this purpose
[7]. Recently, authors have increasingly paid attention to the issue of representativeness of data in
samples, but the current state suggests that data sets have gaps in the representation of gender and
race, and the complex nature of demographic variables makes classification difficult and
inconsistent. Thus, the question of representativeness of data in sets that include people with
disabilities and the elderly is considered. The authors recommend increasing representativeness by
adding samples for underrepresented groups, including by collecting additional data or using
synthetic data methods to improve representation of minorities and people with disabilities.
In the article [8], the authors raise the important problem of sample representativeness in the
context of machine learning and artificial intelligence, emphasizing the need for accurate
representation of population data. The main strategy that the authors propose to achieve high
quality models is the use of stratified samples, which allow to reduce the variability between
subgroups and accurately reflect the proportions between different categories in the population.
The authors of the study [9] consider biases arising both from class imbalances in the data and
from sensitive (protected) characteristics such as race or gender. The approach increases model
accuracy by balancing classes and reduces dependence on sensitive features, which improves group
fairness.
IBM researchers have developed an open-source AI Fairness 360 toolkit for evaluating and
reducing discrimination in machine learning models [10]. The main purpose of the toolkit is to
detect bias based on attributes such as race, gender or age, and to provide methods for
representation of all given social groups at different stages of model development.
The article [11] highlights the problem of intersectional biases in natural language processing
(NLP) models, namely the unrepresentative and biased representation of different groups of people
in textual datasets. The results showed that although existing debiasing methods (for example, for
BERT or RoBERTa) preserve the predictive accuracy of the models well, their ability to reduce
intersectional biases is limited.
The authors of [12] propose a specialized model of machine learning to detect and minimize
bias in textual data, in particular, in news articles. The authors claim that their approach is
effective because of deep models and transformative architectures that are able to detect and
correct biases at different stages of machine learning.
The article [13] presents the problem of gender bias in natural language processing models,
solving it using two main approaches: statistical and causal fairness. Researchers use techniques
such as counterfactual data augmentation for causal debiasing, as well as resampling and revaging
methods for statistical debiasing. The results showed that the combination of these techniques
allows for significant bias in the models by both statistical and causal metrics.
Article [14] is devoted to solving the problem of intersectional bias in the predictions of
machine learning models, in particular deep neural networks. Researchers propose a new method
based on the Apriori algorithm for automatically detecting biased subgroups in data. It allows
efficient generation of frequent subgroups and calculation of fairness metrics for them.
In [15], the authors identify and classify bias in natural language processing using transformer
models such as BERT. The authors explore different ways to identify bias, including identifying
social characteristics such as gender, race, religion, and sexual orientation.
The study [16] examines the problem of cyberbullying, which is a threat to people based on
different characteristics, such as religion, age, ethnicity, and gender. The data set used by the
authors has been modified with ethical considerations in mind, which ensures responsible AI.
The cited works show that the formation of representative and unbiased samples is a relevant
research area, however, most of the works are devoted either to the detection of unbiasedness or to
the analysis of the representativeness or unbiasedness of data samples, however, data samples must
be modified to achieve compliance with FATE principles.
So, summarizing, it is possible to highlight the features of the modern approach, which is
applied to the development of AI models (Fig. 1). However, this approach does not take into
account existing ethical principles and non-discriminatory, representative presentation of existing
population subgroups, which should be applied to obtain AI models.
Figure 1: An existing approach to training AI models.
The purpose of the work is to ensure compliance with the ethical aspects (gender, religious, age,
etc.) of the FATE principle of justice [5] for educational datasets, which consists in creating a
method of analysis and formation of representative (according to the specified aspects) text
samples of data. To achieve the specified goal, it is necessary to propose a method that will
implement the following research tasks:
• to develop an approach to the analysis and formation of relevant representative datasets
according to the principle of fairness of FATE for subject areas.
• to investigate the effectiveness of the proposed approach, by using it for the applied
analysis of the text dataset and bringing it to a representative view according to the aspects
of the FATE principle of justice: gender, age and religion.
3. Method for analysis and formation of representative text datasets
In contrast to the existing approach to training AI models (see Fig. 1), the study proposes a new
approach (Fig. 2), which will ensure the representativeness and ethical correctness of the datasets
used for training AI models.
Figure 2: An approach to the formation of ethically representative datasets.
In order to implement the proposed approach, we will present: the information model and
presentation of the task of forming representative samples of text data as an optimization task;
steps of the method of analysis and formation of representative samples; a way to obtain a typical
ML model for the ethical aspect; description of the composition of the datasets for the study.
3.1. Information model
The problem of obtaining a representative, ethically unbiased text dataset can be presented in the
framework of an information model of the following form:
{𝐷𝐷𝐷𝐷, 𝐷𝐷𝐷𝐷ʹ, 𝐶𝐶, 𝐴𝐴, 𝑀𝑀, 𝐹𝐹}, (1)
where DS is the text dataset for analysis and correction, DSʹ is the text dataset after correction, C
is the set of classes of the subject domain of the dataset, A is the set of ethical aspects, M is the set
of trained machine learning models (separate for each ethical aspect), F is the objective function
minimizing the deviation between current and desired ratios for all ethical aspects.
In (1), the initial dataset DS and the corrected dataset DSʹ can be represented as:
{𝐷𝐷𝐷𝐷 = {𝐷𝐷 ∪ 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀}, (2)
DSʹ = {Dʹ ∪ Metadataʹ}, (3)
where D is the set of elements of the DS dataset, Metadata is the set of metadata of the DS
dataset, Dʹ is the set of elements of the DSʹ dataset, Metadataʹ is the set of metadata of the
DSʹ dataset.
Each element of the set of elements of the dataset D in (2) and each element of the set of
elements of the dataset Dʹ in (3) is a tuple of the following form:
𝑑𝑑 = 𝑑𝑑ʹ = (𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡, 𝑐𝑐𝑥𝑥 , 𝐴𝐴𝐶𝐶𝑥𝑥 ), (4)
where the attribute text is the textual content of element d or dʹ; cx is the class of the subject
area of the dataset to which the element belongs, cx ∈ C; ACx is a set of classes of dataset element
belonging to ethical aspects.
Thus, in (4) cx and ACx are the marking (marking) of the content of the text element.
In (4), the set of classes of membership of the dataset element DS or DSʹ in (1) to the ethical
aspects Ax is presented in the form of a tuple:
𝐴𝐴𝑥𝑥 = (𝑎𝑎1𝑥𝑥 , 𝑎𝑎2𝑥𝑥 , … , 𝑎𝑎𝑘𝑘𝑘𝑘 , ), (5)
where ax – classes of element belonging to ethical aspects; k is the number of ethical aspects to
be analyzed, k = |Ax|.
At the same time, in (5) according to (1) Ax ⊂ A, and classes of dataset elements belonging to
ethical aspects are elements of the corresponding sets, unique for each of the ethical aspects:
𝑎𝑎 1,𝑥𝑥 ∈ 𝐴𝐴1 , 𝑎𝑎2,𝑥𝑥 ∈ 𝐴𝐴2 , … , 𝑎𝑎𝑘𝑘,𝑥𝑥 ∈ 𝐴𝐴𝑘𝑘 , (6)
𝐴𝐴1 ∪ 𝐴𝐴2 ∪ … ∪ 𝐴𝐴𝑘𝑘 = 𝐴𝐴 (7)
The Metadata set of the DS dataset in (2) includes:
𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝐷𝐷𝐷𝐷 = {𝑛𝑛𝐷𝐷𝐷𝐷 , 𝐴𝐴𝐴𝐴𝐷𝐷𝐷𝐷 , 𝐴𝐴𝑇𝑇𝐷𝐷𝐷𝐷 , 𝑛𝑛ʹ𝐷𝐷𝐷𝐷 , 𝐴𝐴𝑁𝑁ʹ𝐷𝐷𝐷𝐷 , 𝐴𝐴𝐴𝐴ʹ𝐷𝐷𝐷𝐷 }, (8)
where nDS is the number of elements in D, nDS = |D|; ANDS is the set of quantities of dataset
elements belonging to each class of each ethical aspect from Ax; ATDS is the set of available
proportions of items for each class relative to the total for each ethical aspect from Ax, nʹDS is the
target number of elements in Dʹ; ANʹDS is the set of target quantities of dataset elements belonging
to each class of each ethical aspect from Ax; ATʹDS is the set of target proportions of elements for
each class relative to the total amount for each ethical aspect from Ax.
At the same time, in (8), each element anDS,i of the set ANDS corresponds to a separate i-th
ethical aspect and is represented by a tuple of the following form:
𝑎𝑎𝑎𝑎𝐷𝐷𝐷𝐷,𝑖𝑖 = (𝑛𝑛𝐷𝐷𝐷𝐷,𝑖𝑖,1 , 𝑛𝑛𝐷𝐷𝐷𝐷,𝑖𝑖,2 , … , 𝑛𝑛𝐷𝐷𝐷𝐷,𝑖𝑖,𝑗𝑗 , … , 𝑛𝑛𝐷𝐷𝐷𝐷,𝑖𝑖,𝑘𝑘 ), (9)
where nDS,i,1 is the number of elements in the dataset of the 1st class of the i-th ethical aspect,
nDS,i,2 is the number of elements in the dataset of the 2nd class of the i-th ethical aspect, nDS,i,j is the
number of elements in the dataset of the j-th class of the i-th ethical aspect, k is the number of
classes of the i-th ethical aspect.
Similarly to (9), in (8) the proportions of the elements atDS,i of the i-th ethical aspect are
represented by a tuple of the following form:
𝑎𝑎𝑎𝑎𝐷𝐷𝐷𝐷,𝑖𝑖 = (𝑡𝑡𝐷𝐷𝐷𝐷,𝑖𝑖,1 , 𝑡𝑡𝐷𝐷𝐷𝐷,𝑖𝑖,2 , … , 𝑡𝑡𝐷𝐷𝐷𝐷,𝑖𝑖,𝑗𝑗 , … , 𝑡𝑡𝐷𝐷𝐷𝐷,𝑖𝑖,𝑘𝑘 ), (10)
where tDS,i,1 is the ratio of the number of elements in the dataset of the 1st class of the i-th
ethical aspect to the total number of elements in the dataset, tDS,i,2 is the ratio of the number of
elements in the dataset of the 2nd class of the i-th ethical aspect to the total number of elements in
the dataset, tDS,i,j is the ratio of the number of elements in the dataset of the i-th class of the
i-th ethical aspect to the total number of elements in the dataset.
At the same time, for the values (9) and (10) in accordance with (8) for each i-th ethical aspect,
the equality holds:
𝑛𝑛𝐷𝐷𝐷𝐷,𝑖𝑖,1 + 𝑛𝑛𝐷𝐷𝐷𝐷,𝑖𝑖,2 + … + 𝑛𝑛𝐷𝐷𝐷𝐷,𝑖𝑖,𝑘𝑘 = 𝑛𝑛𝐷𝐷𝐷𝐷 , (11)
𝑡𝑡𝐷𝐷𝐷𝐷,𝑖𝑖,1 + 𝑡𝑡𝐷𝐷𝐷𝐷,𝑖𝑖,2 + … + 𝑡𝑡𝐷𝐷𝐷𝐷,𝑖𝑖,𝑘𝑘 = 1. (12)
In contrast to (8), the set of Metadataʹ of the DSʹ dataset in (3) includes:
𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀ʹ𝐷𝐷𝐷𝐷 = {𝑛𝑛ʹʹ𝐷𝐷𝐷𝐷 , 𝐴𝐴𝐴𝐴ʹʹ𝐷𝐷𝐷𝐷 , 𝐴𝐴𝐴𝐴ʹʹ𝐷𝐷𝐷𝐷 }, (13)
where nʹʹDS is actually the number of elements in Dʹ obtained as a result of adjustment,
nʹʹDS = |D|; ANʹʹDS is the set actually obtained as a result of adjusting the quantities of dataset
elements belonging to each class of each ethical aspect from Ax; ATʹʹDS is the set actually obtained
as a result of adjusting the proportions of elements for each class relative to the total number for
each ethical aspect from Ax.
Thus, in (8) and (13), (9) and (11) hold for ANʹDS and ANʹʹDS, and (10) and (12) hold for
ATʹDS and ATʹʹDS .
Thus, according to (4), (6) and (7), the text dataset D has the number of elements
n = nDS = |D| and can be presented in the form:
𝐷𝐷 = {𝑑𝑑1 , 𝑑𝑑2 , … , 𝑑𝑑𝑛𝑛 , }, 𝑑𝑑𝑖𝑖 = (𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑖𝑖 , 𝑐𝑐𝑖𝑖 , 𝐴𝐴1 , 𝐴𝐴2 , … , 𝐴𝐴𝑚𝑚 ), 𝑖𝑖 = ���������
1, … , 𝑛𝑛 (14)
where C = {c1, c2, …, ck}, where k is the number of classes of dataset D, m is the number of ethical
aspects.
According to (6) – (10), the solution of the problem is aimed at obtaining the dataset Dʹ, which
contains the total number of elements nʹ = nʹDS = |Dʹ|, quantitatively balanced according to the
ethical aspects Аi from the set of ethical aspects А:
𝐴𝐴 = {𝐴𝐴1 , 𝐴𝐴2 , … , 𝐴𝐴𝑚𝑚 }, 𝐴𝐴𝑖𝑖 = �𝐶𝐶𝑖𝑖 , 𝑇𝑇𝑖𝑖𝑖𝑖 �, 𝑖𝑖 = ���������
1, … , 𝑚𝑚, (15)
where each aspect Ai contains classes Сi and target proportions of classes Tij or each element of
class С; С is the set of classes of the ethical aspect Ai , C = {c1, c2, …,cj}; j is the number of classes of
the ethical aspect of Ai.
To balance the dataset for each ethical aspect, it is necessary to use trained or train an
appropriate number of classifier models, which can be as deep learning models, for example, BERT,
LSTM, GRU, as well as machine learning models of Logistic Regression, Naive Bayes, Support
Vector Machines, k-Nearest Neighbors etc. [17], and according to (1), the set of trained models of
classifiers M is presented in the form:
𝑀𝑀 = {𝑀𝑀1 , 𝑀𝑀2 , … , 𝑀𝑀𝑚𝑚 }, m = |D|. (16)
Thus, within the framework of the proposed information model, it is necessary to perform the
transformation D ⇒ Dʹ with the condition of maximal correspondence nʹʹDS → nʹDS,
ANʹʹDS → ANʹDS та ATʹʹDS → ATʹDS.
3.2. Idea of the approach
The study proposes to reduce the task of building a representative, ethically unbiased dataset to the
task of multi-criteria optimization. The optimization task consists in minimizing the deviation
between the current and desired class ratios, taking into account the limitations on the number of
samples in the classes and the possibilities of generating synthetic data.
Input data: textual dataset DS, set of ethical aspects A, requirements for representative
distribution DSʹ.
The goal of the problem: to create a representative sample for all ethical aspects that achieves the
target class proportions for each ethical aspect D ⇒ Dʹ.
Variables: xij – number of samples of class Cj in aspect Ai after sequestration and augmentation.
The objective function F is the minimization of the deviation between the current and desired ratios
for all ethical aspects simultaneously, taking into account constraints (18) – (21):
𝑚𝑚 𝑛𝑛𝑖𝑖
𝑥𝑥𝑖𝑖𝑖𝑖
𝐹𝐹 = 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 � � � − 𝑇𝑇𝑖𝑖𝑖𝑖 �. (17)
𝑛𝑛ʹ
𝑖𝑖−1 𝑗𝑗−1
Limitations of the task:
1) the sum of all class samples within one aspect is equal to the target number of samples for
this aspect (4):
𝑛𝑛𝑖𝑖
(18)
� 𝑥𝑥𝑖𝑖𝑖𝑖 = 𝑛𝑛ʹ, ∀𝑖𝑖 ∈ {1,2, … , 𝑚𝑚},
𝑗𝑗=1
where ni is the number of classes in the aspectAi;
2) the number of samples for each class should correspond to the target proportion of classes:
𝑥𝑥𝑖𝑖𝑖𝑖
≈ 𝑇𝑇𝑖𝑖𝑖𝑖 , ∀𝑖𝑖 ∈ {1,2, … , 𝑚𝑚}, ∀𝑗𝑗 ∈ {1,2, … , 𝑛𝑛𝑖𝑖 }; (19)
𝑛𝑛ʹ
3) the estimated number of samples cannot be negative:
𝑥𝑥𝑖𝑖𝑖𝑖 ≥ 0, ∀𝑖𝑖 ∈ {1,2, … , 𝑚𝑚}, ∀𝑗𝑗 ∈ {1,2, … , 𝑛𝑛𝑖𝑖 }; (20)
4) the ability to add new samples should match the ability to generate new data for each class
and aspect:
𝑥𝑥𝑖𝑖𝑖𝑖 ≤ 𝐺𝐺𝑖𝑖𝑖𝑖 , ∀𝑖𝑖 ∈ {1,2, … , 𝑚𝑚}, ∀𝑗𝑗 ∈ {1,2, … , 𝑛𝑛𝑖𝑖 }, (21)
where Gij is the maximum possible number of samples of class Cj in aspect Ai, that can be added.
Based on the set optimization task of forming a representative dataset (17), we present the
steps of the method of analysis and formation of representative samples of text data.
3.3. Main steps of method
Method for analysis and formation of representative text datasets is presented in the form of
three consecutive stages: preprocessing, analysis of representativeness according to ethical
aspects and representative adjustment of dataset. Each stage consists of its own steps, which
are shown in Figure 3.
The input data of the method is the dataset DS for analysis, which according to (2) and (8)
contains the target number of nʹDS elements, the set of ethical aspects A with subsets of classes,
the target proportions of ATDS classes and the number of ANʹDS elements in the classes of
ethical aspects, respectively, the trained set of models М for each ethical aspect from A, which
uses balanced samples for each ethical aspect for training.
At stage 1, a sample of text data in D ⊂ DS is pre-processed, namely, the removal of non-
informative text fragments such as punctuation marks, numbers and special characters [18].
Removal of emoticons is not performed, as in many cases including emoticons in the analysis
improves the accuracy of machine learning models used to classify texts based on emotional or
mood content [19]. Incorrect records (empty, uninformative, etc.) are also deleted.
At stage 2, an analysis of the representativeness of the sample of textual data is carried out,
taking into account ethical aspects. First, it is necessary to vectorize and classify each element
∀d ∈ D of the data sample using separate machine learning models m ∈ M for each of the
ethical aspects Ai ∈ A. The existing proportions of ANDS and ATDS classes for each of the ethical
aspects are determined. The amount of shortage or excess of elements of each class for each of
the ethical aspects is calculated. After that, the sufficiency of the data in the sample for
augmentation is analyzed (minimum availability of samples of the relevant classes, etc.).
Stage 3 involves a representative adjustment of the data sample to take into account ethical
considerations. Adjustments include removing and adding.
The deletion operation is performed to remove redundant elements of each class for each of the
ethical aspects with minimal damage to other distributions, for which the optimization problem of
selecting redundant elements in the framework of (17), which should be removed to achieve the
target proportions of classes, is solved.
The add operation is performed to create new items using one of the known methods, for
example, the SMOTE method [20]. Requirements are created in the form of the necessary
combination of classes of each of the ethical aspects for each new element, for which the
optimization problem of forming requirements for the missing elements is solved within the
framework of (17).
Figure 3: Steps of method for analysis and formation of representative text datasets.
The output data of the method is a text dataset Dʹ ⊂ DSʹ, which has the required volume nʹDS and
is balanced according to the required proportions ATʹDS according to the selected ethical aspects Ax
⊂ A.
The steps of the method of analyzing and generating representative samples of text data will
allow you to generate text samples that are non-discriminatory and unbiased and reflect a
proportional representation of the sample samples to the actual demographic subgroups of the
population, which will affect the accuracy and transparency of training machine learning models
for solving various problems.
4. Results and discussion
4.1. Datasets for research
To test the method of analysis and formation of representative samples of textual data, an input
dataset was formed based on two datasets "Cyberbullying Classification" [3] and "Cyberbully
Detection Dataset" [4]. The "Cyberbullying Classification" dataset contains 46,017 tweets, which
are labeled by types of cyberbullying into 6 classes. The "Cyberbully Detection Dataset" contains
99,989 tweets, which is also labeled by type of cyberbullying. Both datasets are unlabeled for
gender, age group, religion, and ethnicity of the message author.
To train machine learning models, which will be used to label the input dataset, datasets were used
on the example of three ethical aspects of the principle of justice of gender, age and religion.
The English-language dataset "Tweet Files for Gender Guessing" [21], which contains 34,146
unique text entries, which are divided into two classes: female and male, with 17,073 entries in each
class, was used to train ML based on the ethical aspect of the gender of the author of the message.
On the basis of the English-language dataset "CyberBullying Detection Dataset" [22], which
contains 20109 test samples, a sample was created for training the classifier and marking the input
dataset according to the religious ethical aspect. The dataset in Italian "TAG-it Dataset
Distribution" [23] was translated into English and used to bring to a representative view the
working dataset by age and contains 21,948 text messages divided into age classes: 0-19, 20-29, 30-
39, 40-49, 50-100 years old.
Since the classes in the given datasets are not balanced and have a different number of samples,
which will negatively affect the quality of training of machine learning models, all classes in the
datasets were balanced in terms of number. The final number of samples in each class of training
samples for ML training according to ethical aspects is shown in Fig. 4.
Figure 4: Classes and number of samples in ML training datasets for ethical aspects.
As a result of the work on creating training samples, datasets balanced by the number of text
messages in classes were obtained. Such datasets will make it possible to correctly assess the
representativeness of working text datasets.
4.2. Software for research
To study the effectiveness of the method of analysis and formation of a representative sample of
text data, a software implementation was created using the Python programming language. The
tensorflow library (https://www.tensorflow.org/) was used to classify the input dataset on
cyberbullying based on gender, age, and religion. In Fig. 5 shows an example of classification based
on the religious basis of the FATE-principle of justice.
Figure 5: Developed software for classification of dataset by religious ethical aspect.
To form a set of trained ethical machine learning models, which are separate for each ethical
aspect, various classifier models were analyzed, and to select the best of them, their quality was
evaluated by statistical indicators, such as Accuracy, Precision, Recall, and F1-score [24]. Both deep
learning models, such as BERT, GPT, LSTM, GRU, etc., and classifiers such as Logistic Regression,
Naive Bayes, Support Vector Machines k-Nearest Neighbors, etc., were studied [25]. After that, the
classifier is trained on the selected ML model on the annotated dataset for the ethical aspect.
As a result, different architectures were chosen as classifiers: FastForest classifiers, SVM and
LSTM, BERT deep learning models [26]. Thus, machine learning models such as FastForest, SVM,
LSTM, and BERT are effective tools for solving text classification tasks, including determining a
person's gender, religion, and age based on user text posts. Classical approaches such as FastForest
and SVM have also demonstrated their effectiveness in text classification. FastForest works efficiently
with large datasets and prevents overtraining. SVM, in turn, is known for its ability to work with
high-dimensional data, which is especially useful for text classification, where each word or phrase
can be represented as a separate feature [27]. Deep learning models, such as LSTM and BERT, are
able to recognize complex patterns in text sequences, preserving the context at all stages of analysis
[28]. A distinctive feature of LSTM is its ability to retain information about previous parts of the text,
which makes this model effective for complex classification tasks where the overall context of the
message is important. Studies have shown that such a model can achieve an accuracy of up to 92% in
text classification tasks [29]. The BERT model, in turn, is characterized by the ability to analyze the
text in two directions, that is, to take into account both the previous and subsequent context of words
[30].
4.3. Analysis of research results
To analyze and form a representative sample of text data for the target proportions of classes, to
form a representative sample of text data by age and gender, the population of Ukraine was
taken. According to the M. V. Ptukh Institute of Demography and Social Research of the National
Academy of Sciences of Ukraine (https://idss.org.ua/forecasts/nation_pop_proj), as of July 2023,
the total population of Ukraine is estimated at 3,559,6216 people. The following number of people
is represented in each age subgroup: age group 0-19 years 6,659,068 people, 20-29 years 3,623,143
people, 30-39 years - 6,022,345 people, 40-49 years - 5,431,140 people, 50 -100 years - 13,860,520
people. Regarding the gender structure of the population of Ukraine in 2023: 1,695,1527 are
women, and 1,864,689 are men (idss.org.ua). Note that within the scope of this work, the
cisgender group is considered in the analysis of the gender ethical aspect.
To study the effectiveness of the method of analysis and formation of a representative
selection of text data described in the work, several machine learning models were trained. The
results of calculating static metrics such as Accuracy, Precision, Recall and F1-score [24] of
machine learning models for the gender, age and religious ethical aspects are shown in Table 1.
Table 1
Statistical metrics Accuracy, Precision, Recall and F1-score of machine learning models by gender,
age and religious ethical aspects
ML model Accuracy Precision Recall F1-score
Gender ethical aspect
FastForest 0.630 0.640 0.600 0.620
SVM 0.580 0.580 0.580 0.580
LSTM 0.70 0.770 0.670 0.720
BERT 0.690 0.640 0.710 0.670
Age ethical aspect
FastForest 0.535 0.542 0.504 0.504
SVM 0.815 0.770 0.779 0.770
LSTM 0.590 0.600 0.560 0.580
BERT 0.580 0.430 0.450 0.440
Religious ethical aspect
FastForest 0.775 0.800 0.762 0.780
SVM 0.825 0.850 0.810 0.829
LSTM 0.850 0.880 0.830 0.854
BERT 0.910 0.980 0.74 0 0.840
For different classes, different levels of linear resolution were obtained: according to religion
using the BERT classifier, which showed the best result of the trained machine learning models for
the task of classifying text samples according to the religious ethical aspect, the data turned out to
be well separated, according to gender using the LSTM classifier, which showed the best
performance compared to other models, the data turned out to be moderately separable, and
according to age, using the SVM classifier, it was poorly separable.
In addition, it was found that the dataset is not representative, because the classes of various
ethical aspects have a number of text samples that do not correspond to the proportions of the
demographic subgroups of the population of Ukraine, thus they need balancing to acquire a
representative appearance.
Therefore, according to the steps of the method of analysis and formation of a representative
sample of text data, a sample of text data needs data augmentation to form a representative sample.
For this, it is necessary to solve the optimization problem, for the correct removal of redundant
elements of each class according to each of the ethical aspects, with further augmentation of the
data sample to the target requirements (number of elements and proportions of classes).
Table 2 presents the percentages of samples by age in the sample of textual data and individuals
of the population in age-demographic subgroups, and also calculates the new distribution of the
sample classes if only one ethical aspect - age - was taken into account.
Table 2
Percentage ratios of samples by age in the sample of text data and individuals of the population in
age demographic subgroups, %
Percentage of Deviation of
Percentage of New
Age population in text samples Deviation from
samples by distribution
demographic age from representative
age in text of sampling
subgroups demographic subgroups of distribution
dataset classes
subgroups population
0-19 years 48.23% 18.71% 29.52% 18.75% 0.04%
20-29 years 2.11% 10.17% 8.06% 10.15% 0.02%
30-39 years 25.28% 16.92% 8.36% 16.87% 0.03%
40-49 years 13.60% 15.26% 1.66% 15.28% 0.02%
50-100 years 10.78% 38.94% 28.16% 38.95% 0.01%
Table 3 presents the percentages of samples by gender in the sample of textual data and
individuals of the population in gender demographic subgroups, and also calculates the new
distribution of sample classes if only one ethical aspect - gender - was taken into account.
The deviation of the sample distributions by classes of the age-ethical aspect of the dataset,
transformed according to the created method, from the ideal representative distribution was
obtained: minimum 0.01%, maximum 0.04%, average 0.02%, and for the gender ethical aspect:
minimum 0.03%, maximum 0.03%, average 0.03 %.
Table 3
Percentages of samples by gender in the sample of text data and individuals of the population in
gender demographic subgroups, %
Percentage of Deviation of
Percentage of New Deviation
Gender population in text samples
samples by distribution from
demographic gender from
gender in of sampling representative
subgroups demographic subgroups of
text dataset classes distribution
subgroups population
Men 58.94% 43.28% 15.67% 43.25% 0.03%
Women 41.06% 56.72% 15.67% 56.75% 0.03%
However, the optimization task of forming a representative sample of textual data is a
multi-criteria one, in which the criteria are the formation of a sample based on age and gender
ethical aspects, so the goal is to minimize the deviation between the current and desired class
ratios, taking into account the limitations on the number of samples and the possibility of
generating new data. As a result of solving the optimization problem for the formation of a
representative sample by age and gender ethical aspects on the example of demographic
subgroups of the population of Ukraine, a representative sample of text data was obtained by
augmentation, the balance of classes of which is presented in Table 4, Fig. 6 and Fig. 7.
Table 4
Distribution of samples in the formed representative sample after data augmentation as a result of
solving a multi-criteria optimization problem
Age
demographic 0-19 years 20-29 years 30-39 years 40-49 years 50-100 years
subgroups
Percentage ratio of demographic groups by gender and age in the population of Ukraine
Men 9.67% 5.64% 8.96% 7.79% 15.56%
Women 9.04% 4.53% 7.96% 7.47% 23.38%
Percentage ratio of demographic groups by gender and age in the text sample
Men 9.65% 5.62% 8.94% 7.80% 15.57%
Women 9.05% 4.57% 7.97% 7.45% 23.38%
The resulting deviation from a representative distribution
Men 0.02% 0.02% 0.02% 0.01% 0.02%
Women 0.01% 0.04% 0.01% 0.02% 0.00%
The deviation of the sample distributions by classes of age and gender ethical aspects of the
dataset simultaneously, transformed according to the created method, from the ideal representative
distribution was obtained: minimum 0.00%, maximum 0.04%, average 0.02%.
So, as a result of performing the steps of the analysis method and forming representative
samples of text data, a text sample was formed, which is non-discriminatory and unbiased and
reflects the representation of sample samples proportional to the real demographic subgroups of
the population of Ukraine.
Figure 6: The balance of the distribution of the input dataset according to the age-ethical aspect of
the FATE-principle of justice.
Figure 7: The balance of the distribution of the input dataset according to the gender ethical aspect
of the FATE-principle of justice.
5. Conclusion
Thus, the goal of the study was achieved through the development of the method for analysis and
formation of representative text datasets, designed for the analysis and formation of representative
text samples of data according to the principle of fairness of FATE for subject areas.
To investigate the effectiveness of the analysis method and the formation of a
representative presentation of the text dataset, software was created that uses machine
learning models to classify texts according to various ethical aspects - age, gender, religion,
ethnicity, etc. Thus, to classify the text samples in the sample according to the age-ethical
aspect, SVM was used, LSTM was used for gender, and BERT was used for religious ones,
which are the best indicators of statistical metrics.
As a result of the practical application of the developed method, it was established that the
available dataset is not representative compared to the objective data of demographic statistics,
so a multi-criteria optimization problem was solved and the dataset was transformed into a
representative one in terms of age and gender ethical aspects. The obtained deviations of the
sample distributions by classes of ethical aspects of the dataset transformed according to the
created method from the ideal representative distribution were: minimum 0.00%, maximum
0.04%, average 0.02%, under the conditions of the initial volume of the dataset 47,692 elements,
the minimum initial number of samples in the class 1007 elements, the maximum initial
number of samples in the class is 28,112 elements. The studied efficiency proves that the
developed method allows performing the analysis of the representativeness of text datasets and
bringing them to a representative form according to various aspects of the FATE fairness
principle.
The obtained results contribute to improvement of representativeness of text datasets and
fair and unbiased representation of demographic groups in them, which increases trust in
decisions made by artificial intelligence, and complies with goals SDG3 (good health and well-
being), SDG4 (quality education) and SDG16 (peace, justice, and strong institutions).
Further plans for improving the method of analysis and formation of representative samples
of text data are the formation of not only a non-discriminatory sample by the number of
samples, but also the search and removal of samples of text samples that contain a biased
attitude towards representatives of various demographic subgroups, according to the ethical
aspects of the FATE-principle of justice.
Also, the prospects for further research are the use of the developed method for adjusting
textual datasets of subject areas and their use for solving applied problems, such as detection
and classification of cyberbullying, analysis of the emotional tonality of messages, detection of
the physical and mental state of users based on their posts, etc. Detecting performance gains
from using ethically balanced text datasets will provide feedback for improving the developed
method.
Declaration on Generative AI
During the preparation of this work, the authors used Grammarly in order to: grammar and
spelling check; DeepL Translate in order to: some phrases translation into English. After using
these tools/services, the authors reviewed and edited the content as needed and take full
responsibility for the publication’s content.
References
[1] M. Shah, N. Sureja, A Comprehensive Review of Bias in Deep Learning Models: Methods,
Impacts, and Future Directions, Arch Computat Methods Eng (2024). doi:10.1007/s11831-
024-10134-2
[2] Y. Yusyn, N. Rybachok, dictionary-based deterministic method of generation of text
CORPORA, Computer systems and information technologies 3 (2024) 67–73..
doi:10.31891/csit-2024-3-9.
[3] Kaggle.com, Cyberbullying Classification, 2021. URL:
https://www.kaggle.com/datasets/andrewmvd/cyberbullying-
classification?resource=download
[4] Kaggle.com, CyberBullying Detection Dataset, 2024. URL:
https://www.kaggle.com/datasets/sayankr007/cyber-bullying-data-for-multi-label-
classification
[5] B. C. Stahl, D. Eke, The ethics of ChatGPT – Exploring the ethical issues of an emerging
technology, International Journal of Information Management 74 (2024), p. 102700.
doi:10.1016/j.ijinfomgt.2023.102700
[6] E. Manziuk, I. Krak, O. Barmak, O. Mazurets, V. Kuznetsov, O. Pylypiak, Structural
alignment method of conceptual categories of ontology and formalized domain, CEUR
Workshop Proceedings 3003 (2021) pp. 11–22
[7] O. Barmak, O. Mazurets, I. Krak, A. Kulias, A. Smolarz, L. Azarova, K. Gromaszek, S.
Smailova, Information technology for creation of semantic structure of educational
materials, Proceedings of SPIE – The International Society for Optical Engineering 11176
(2019), pp. 147–156. doi:10.1117/12.2537064
[8] L. K. H. Clemmensen, D. K. Rune, Data Representativity for Machine Learning and AI
Systems, 2022. URL: https://ar5iv.labs.arxiv.org/html/2203.04706
[9] D. Dablain, B. Krawczyk, N. Chawla, Towards a holistic view of bias in machine learning:
bridging algorithmic fairness and imbalanced learning. Discov Data 2, 4 (2024).
doi:10.1007/s44248-024-00007-1
[10] R. K. E. Bellamy, AI Fairness 360: An extensible toolkit for detecting and mitigating
algorithmic bias, IBM Journal of Research and Development, volume 63, 4/5, (2019) pp. 1-
15. doi:10.1147/JRD.2019.2942287
[11] J. Lalor, Y. Yang, K. Smith, N. Forsgren, A. Abbasi, Benchmarking Intersectional Biases in
NLP, Proceedings of the 2022 Conference of the North American Chapter of the
Association for Computa-tional Linguistics: Human Language Technologies, Association
for Computational Linguistics (2022), pp. 3598–3609. doi:10.18653/v1/2022.naacl-main.263
[12] S. Raza, D.J. Reji, C. Ding, Dbias: detecting biases and ensuring fairness in news articles.
International Journal of Data Science and Analytics, volume 17 (2024), pp. 39–59.
doi:10.1007/s41060-022-00359-4
[13] H. Chen, Y. Ji, D. Evans. Addressing Both Statistical and Causal Gender Fairness in NLP
Models. In Findings of the Association for Computational Linguistics: NAACL 2024,
Association for Computational Linguistics (2024), pp. 561–582. doi:
10.48550/arXiv.2404.00463
[14] K. Zhou, J. Wen, N. Yang, D. Yuan, Q. Lu, H. Chen, Fairpriori: Improving Biased Subgroup
Discovery for Deep Neural Network Fairness (2024). doi:10.48550/arXiv.2407.01595
[15] A. S. Evans, H. Moniz, L. Coheur, A Study on Bias De-tection and Classification in Natural
Language Processing (2024). doi:10.48550/arXiv.2408.07479
[16] A. Orelaja, C. Ejiofor, S. Sarpong, S. Imakuh, C. Bassey, I. Opara, J. N. A. Tettey, O.
Akinola, Attribute-Specific Cyberbullying Detection Using Artificial Intelligence, Journal
of Electronic & Information Systems, volume 6(1) (2024), pp. 10–21.
doi:10.30564/jeis.v6i1.6206
[17] M. Zulqarnain, R. Sheikh, S. Hussain, M. Sajid, S. N. Abbas, M. Majid, U. Ullah, Text
Classification Using Deep Learning Models: A Comparative Review, Cloud Computing and
Data Science (2024), pp. 80-96. doi: 10.1007/s10115-023-01856-z
[18] O. Zalutska, M. Molchanova, O. Sobko, O. Mazurets, O. Pasichnyk, O. Barmak, I. Krak,
Method for Sentiment Analysis of Ukrainian-Language Reviews in E-Commerce Using
RoBERTa Neural Network, CEUR Workshop Proceedings 3387 (2023), pp. 344–356.
doi:10.15407/jai2024.02.085
[19] H. Tang, W. Tang, D. Zhu, S. Wang, Y. Wang, L. Wang, EMFSA: Emoji-based multifeature
fusion sentiment analysis. PLoS One. 19(9) (2024). doi: 10.1371/journal.pone.0310715
[20] T. He. S. Wongvorachan, O. Bulut, A comparison of undersampling, oversampling, and
SMOTE methods for dealing with imbalanced classification in educational data mining,
Information, 14(1), 54 (2023). doi:10.3390/info14010054
[21] Kaggle.com, Tweet Files for Gender Guessing, 2019. URL:
https://www.kaggle.com/datasets/aharless/tweet-files-for-gender-guessing
[22] Kaggle.com, CyberBullying Detection Dataset, 2024. URL:
https://www.kaggle.com/datasets/sayankr007/cyber-bullying-data-for-multi-label-
classification
[23] Live.european-language-grid.eu, TAG-it Dataset Distribution, 2024. URL:
https://live.european-language-grid.eu/catalogue/corpus/8112/download/
[24] K. Nazeri, H. Thasarathan, M. Ebrahimi, Edge-informed single image super-resolution,
Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops ,
2019. URL: https://openaccess.thecvf.com/content_ICCVW_2019/html/AIM/Nazeri_Edge-
Informed_Single_Image_Super-Resolution_ICCVW_2019_paper.html
[25] O. Kovalchuk, V. Slobodzian, O. Sobko, M. Molchanova, O. Mazurets, O. Barmak, I. Krak,
N. Savina, Visual Analytics-Based Method for Sentiment Analysis of COVID-19 Ukrainian
Tweets, Lecture Notes on Data Engineering and Communications Technologies 149 (2023)
pp. 591–607. doi:10.1007/978-3-031-16203-9_33
[26] R. Sadeghi, A. Akbari, M.M. Jaziriyan, ExaAUAC: Arabic Twitter user age prediction
corpus based on language and metadata features. Discover Artifcial Intelligence 4, 48
(2024). doi: 10.1007/s44163-024-00145-0
[27] V. Sheth, U. Tripathi, A. Sharma. A Comparative Analysis of Machine Learning
Algorithms for Classification Purpose. Procedia Computer Science (2022), pp. 422-431. doi:
10.1016/j.procs.2022.12.044
[28] Sh. Minaee, E. Cambria, J. Gao. Deep Learning-based Text Classification. ACM Computing
Surveys (CSUR), 54 (2020), pp. 1–40. doi:10.1145/3439726
[29] H. Sa’diah, M. R. Faisal, A. Farmadi, F. Abadi, F. Indriani, M. Alkaff, V. Abdullayev, Gender
Classification on Social Media Messages Using fastText-base Feature Extraction and Long
Short-Term Memory, Journal ofElectronics, Electromedical Engineering, and Medical
Informatics, volume 6(3) (2024), pp. 243–252. doi: 10.35882/jeeemi.v6i3.407
[30] A. Alqahtani, K. H. Ullah, Sh. Alsubai, M. Sha, A. Almadhor, T. Iqbal, S. Abbas. An efficient
approach for textual data classification using deep learning. Frontiers in Computational
Neuroscience, 16 (2022). doi:10.3389/fncom.2022.992296