=Paper=
{{Paper
|id=Vol-3896/paper13
|storemode=property
|title=Improvement of intelligent systems for creating personalized products
|pdfUrl=https://ceur-ws.org/Vol-3896/paper13.pdf
|volume=Vol-3896
|authors=Viktoriia Mytsa,Mykola Riabchykov
|dblpUrl=https://dblp.org/rec/conf/ittap/MytsaR24
}}
==Improvement of intelligent systems for creating personalized products==
https://ceur-ws.org/Vol-3896/paper13.pdf
Improvement of intelligent systems for
creating personalized products
Viktoriia Mytsa1,† and Mykola Riabchykov2,†
1
Khmelnytskyi National University, Instytutska 11, 29016 Khmelnytskyi, Ukraine 2
Lutsk National Technical University, Lvivska 75, 43018 Lutsk, Ukraine
Abstract
The article proposes an information system to ensure the functioning of artificial intelligence systems
in the fashion industry. A model has been developed in the form of a graph, representing the main
characteristics and structural components of a personalized product. This model consists of five main
groups of characteristics (structural, aesthetic, cultural, technological, and economic), which include
114 individual parameters that can vary depending on the specifics of the product and user needs.
Formalized approaches for effective interaction with artificial intelligence systems in the process of
creating personalized products are proposed. Testing was conducted using synthetic data. Multiple
sets of personalized products were created to meet different user needs, and the model's ability to
correctly combine parameters and generate products according to specifications was evaluated.
During field tests, the model was integrated into a recommendation system for clothing selection, and
the results were assessed based on how well they matched user preferences. The test results
demonstrate the potential of the developed system for real-world implementation in the fashion
industry, enabling personalized production of clothing and accessories.
Keywords
artificial intelligence, feature graph, information system, personalized product 1
1. Introduction
Intelligent information systems are a trend in the use of artificial intelligence systems. Research
[1] demonstrates the possibilities of creating a smart environment. The effectiveness of creating
predictive models is considered. Recently, research on Intelligent information systems in the
fashion industry has begun to appear [2]. The need to develop intelligent approaches, models
and structures that combine graphic and text data is noted. The relevance of the implementation
of artificial intelligence systems in the fashion industry is noted in [3]. Emphasis is placed on the
creation of clothing collocation systems, modeling of fashion compatibility, associations
between clothing elements.
∗
ITTAP’2024: 4th International Workshop on Information Technologies: Theoretical and Applied Problems, October 23-
25, 2024, Ternopil, Ukraine, Opole, Poland
∗
Corresponding author.
†
These authors contributed equally
. mitsa_vv@ukr.net (V. Mytsa); mykola.riabchykov@lntu.edu.ua (M. Riabchykov)
0000-0002-5453-9787 (V. Mytsa); 0000-0002-9382-7562 (M. Riabchykov)
© 2023 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
Prospects for the creation of intelligent clothing are considered in the article [4]. The authors
analyze the Hedonic Information Systems Acceptance Model in relation to clothing. The need
for further filling and improvement of this model for the real implementation of artificial
intelligence systems in the field of clothing production was noted. Real examples of the
implementation of intelligent clothing are considered in studies [5, 6].
According to the latest research, one of the urgent issues in the fashion industry is
automating fashion design trend forecasting. The study [7] notes the importance of developing
specialized algorithms and models that take into account the specific features of clothing design.
This study focuses on aspects of color and style, without delving into a detailed analysis of
individual elements of clothing.
The forecasting model of fashionable clothing elements, considered in the article [8], uses the
analysis of coexistence relationships of various clothing elements to determine the key factors
influencing fashion trends.
The considered models require the development and implementation of machine learning
algorithms for effective personalization. Similar algorithms that simplify consumer
personalization are discussed in the article [9]. The practical application of such approaches in e-
commerce systems that use collaborative filtering and personalized rating methods is described
in [10].
According to the study [11], the development of intelligent systems in the fashion industry is
inextricably linked with the introduction of advanced visualization methods based on 3D
technologies. This allows you to create more realistic and interactive representations of fashion
products. The article [12] notes the insufficient amount of data for creating personalized fashion
items. This problem is especially relevant when developing systems that take into account the
individual features of the figure and style of each consumer. Studies [13-14] focus on methods of
creating personalized data based on 3D measurements.
In a number of studies, methods of forecasting fashion trends neural networks are proposed.
These approaches show significant potential in predicting future trends in the fashion industry.
The article [15] discusses algorithms that take into account variations of style, color and other
solutions in the process of forecasting fashion trends. Such algorithms allow analyzing large
volumes of data on current trends, identifying hidden patterns in changing fashion trends, and
predicting future trends with high accuracy.
The generation of exogenous knowledge about forecasting the creation of fashion goods is
considered in the study [16]. Such knowledge involves the creation of a complex database on the
features of fashionable clothing, the development of graphs reflecting the relationships between
various elements of fashion, and the formation of algorithms for trend analysis and forecasting.
Forecasting fashion trends on the basis of style analysis, generation of structures and
analysis of fashion lines determine innovative trends in industry and trade [17]. The article [18]
notes the importance of analytical methods used by business to forecast fashion trends.
The work [19] presents a structure that determines the probability of popularity of clothing
elements. This framework is based on the analysis of various factors influencing consumer
preferences, such as social, economic and cultural changes. The development of such a structure
can significantly improve the methods of forecasting fashion trends, as it allows taking into
account market dynamics and changing consumer preferences in real time.
Research [20] describes the creation of a quasi-autoregressive neural network based on a
complex database and relationships between its elements. Computer vision of fashionable
objects was defined as a fundamental basis for ensuring the effective functioning of artificial
intelligence systems. This approach made it possible to automate the process of recognition and
classification of fashion products, to increase the accuracy of trend forecasting.
The formation of intelligent clothing systems, according to the study [21], also requires the
creation of an extended database on color and other aspects of clothing, as well as determining
the relationships and characteristics of the influence of various elements.
The development of online stylist services requires further digitization and the introduction
of innovative technologies [22]. Special attention is paid to the pairwise ratio of structural
elements, which requires the development of a model that determines the level of detailed
compatibility of various clothing components. To digitize such aspects as material and texture,
color scheme, silhouette and proportions, decorative elements and others, it is necessary to
develop algorithms that take into account the relationships between different characteristics of
clothing.
Research [23] proposes a network model with advanced knowledge for a series of fashion
items. Such a model is able to effectively process complex elements of fashion products, but
needs constant updating and expansion to meet fast-changing fashion trends. Directions for
improvement of such models are given in the study [24] with specific examples.
Collaborative filtering fashion algorithms are a promising direction for introducing artificial
intelligence systems into the fashion industry. Collaborative graph filtering discussed in [25]
uses a convolutional graph network to extend a learnable model. This approach makes it
possible to improve the personalization of recommendations for fashion products, increase the
accuracy of predicting user preferences, and optimize the processes of assortment planning and
inventory management in the fashion industry.
Collaborative filtering is a powerful and efficient tool in forecasting systems, especially in the
context of the fashion industry. This method requires the development of advanced spectral
graph filtering methods to identify the main trends and relationships [26].
Scenario modeling fashion is an important tool for implementing and supporting artificial
intelligence systems in the fashion industry. Effective use of such approaches requires the use of
real data sets [27]. This is because the fashion industry is dynamic and sensitive to many factors,
including social, economic and cultural variables.
Research [28] determines the possibility of finding optimal solutions when creating a
promising system of fashionable clothing and lays the foundation for the development of
algorithms capable of generating innovative designs, taking into account current trends and
forecasts. Methods of optimizing the indicators of a personalized product on specific examples
are considered in [29] and allow to increase the efficiency of production, improve the conformity
of products to consumer expectations.
The article [30] highlights issues related to the possibility of introducing smart clothes when
using artificial intelligence systems. Studies [31-32] are devoted to the issues of introducing
artificial intelligence systems into the work of fashion companies. The need to improve
analytical formalized models for the creation of promising fashion products is noted.
The analysis of the state of the issue determines the urgent need for the development of
analytical methods for describing fashion products with the aim of introducing artificial
intelligence systems into the fashion industry. These methods should provide accurate and
comprehensive characterization of products, which will allow AI systems to generate relevant
and innovative designs.
The work aims to develop a system of analytical models for creating a personalized product
using artificial intelligence systems based on the analysis of the structure and features of a
fashion product.
2. Development of an analytical model of a personalized
product for the implementation of artificial intelligence
systems
The analysis of the structure of a fashion product in the context of the dynamic and sensitive
nature of the fashion industry made it possible to identify the following main groups of
properties and features that characterize its structure and functioning.
Group C0 describes the superstructure of the personalized product. Different structural
components of this group differ depending on how separately the desired product works.
Component C01 describes the functioning of individual products. The C02 component determines
the functioning of the desired product in the style complex.
Structure C1 provides the functional properties of the desired clothing, which determine how
effectively the product performs its tasks, providing comfort and protection of the body. These
properties are closely related to the purpose of clothing, the conditions of its use, and the
individual needs of a person. Separate subgroups of this structure can define:
C1,1 – purpose of the product; C1,2 – anthropometric correspondence; C1,3 – seasonality; C1.4 –
comfort and convenience; C1.5 – practicality; C1.6 – functionality. These subgroups may include
the following elements:
С1,11 – everyday, С1,12 – sports, С1,13 – festive, … ;
С1,21 – size: С1,21,1 – XS, С1,21,2 – S, С1,21,3 – M, С1,21,4 – L, С1,21,5 – XL, … ;
С1,22 – figure type: С1,22,1 – Н-type, С1,22,2 – Х-type, С1,22,3 – А-type, С1,22,4 – V-type; С1,22,5 – О-type … ;
С1,23 – age; С1,24 – gender;
С1,31 – summer, С1,32 – winter, С1,33 – demi-season;
С1,41 – air permeability, С1,42 – elasticity; С1,43 – softness;
С1,51 – wear resistance, С1,52 – ease of care, С1,53 – water repellency; С1,61 –
presence of pockets, С1,62 – fastener, С1,63 – hood, … .
Structure С2 provides the aesthetic properties of the desired clothing, which determine its
appearance, perception and attractiveness. They are key to defining style, fashion direction and
emotional impact on consumers. Separate subgroups of this structure can define:
С2,1 – product style; С2,2 – color; С2,3 – print; С2,4 – decor; С2,5 – silhouette; С2,6 – cut. These
subgroups may include the following elements:
С2,1 – classical, С2,12 – sports, С2,13 – avant-garde, … ;
1
С2,41 – lace, С2,42 – embroidery, С2,43 – application … ;
С2,43,1 – adhesive application, С2,43,2 – sequin application; С2,43,3 – appliqué from rhinestones…;
С2,51 – fitted, С2,52 – free, С2,53 – straight, … ;
С2,61 – sewing sleeve, С2,62 – raglan sleeve, С2,63 – combined sleeve.
Framework С3 provides the cultural and social aspects of desired clothing, which determine
how clothing affects the culture and social identity of an individual or group of people. These
aspects include the influence of cultural norms, traditions, societal standards and social factors
on the choice, style and perception of clothing. The importance of these elements lies in the fact
that clothing not only performs the function of protection, but also becomes a means of self-
expression, communication and formation of social ties. Separate subgroups of this structure
can define:
С3,1 – target audience; С3,2 – cultural context; С3,3– social role; С3,4 – ethical aspects. These
subgroups may include the following elements:
С3,11 – youth, С3,12 – business class, С3,13 – mass market;
С3,21 – national motives, С3,22 – subculture, С3,23 – religiosity;
С3,31 – uniform, С3,32 – special purpose, С3,33 – carnival costume;
С3,32,1 – eco-clothes, С3,32,2 – technological clothing; С3,41 –
ecological production, С3,42 – ethical production.
The С4 structure provides the technological properties of the desired garment, which
determine its production aspects and functionality. Also, this structure forms the basis for
innovations in production, which allows brands to remain competitive and meet the
requirements of modern consumers. Separate subgroups of this structure can define:
С4,1 – fabric type; С4,2 – material texture; С4,3 – innovative materials; С4,4 – smart technologies;
С4,5 – production technologies.
These subgroups may include the following elements:
С4,11 – natural, С4,12 – synthetic, С4,13 – mixed;
С4,21 – smooth, С4,22 – embossed, С4,23 – shiny, С4,24 – matte;
С4,31 – smart fabrics, С4,32 – biomaterials;
С4,41 – fitness trackers, С4,42 – technologies for tracking, С4,43 – monitoring of physiological
indicators, … ;
С4,51 – 3D printing; С4,52 – digital printing; С4,53 – laser cutting.
Structure С5 provides the economic and marketing factors of desirable clothing that
determine its commercial aspects. An important aspect is the integration of data from different
sources, such as social networks, online stores and analytical platforms, which provides a more
accurate modeling of the probability of popularity of specific items of clothing. This allows you
to create adaptive systems that can quickly respond to changes in fashion trends and consumer
preferences. Separate subgroups of this structure can define:
С5,1 – price range; С5,2 – brand; С5,3 – promotion strategy; С5,4 – product life cycle.
These subgroups may include the following elements: С5,11 –
budgetary, С5,12 – average, С5,13 – premium segment;
С5,21 – well-known brands, С5,22 – designer clothes, С5,23 – mass market;
С5,31 – advertising, С5,32 – social media, С5,33 – collaborations; С5,41 – new
collection, С5,42 – sale, С5,43– disposal.
The resulting graph of product connections for the formation of forecasting algorithms based
on artificial intelligence is shown in Fig. 1. This structure is basic and can be improved in the
process of training the AI-system, which allows to increase the effectiveness of predictions in
the dynamic environment of fashion. This, in turn, can significantly influence production
strategies, marketing campaigns and the development of new collections, providing brands with
competitive advantages in the market.
Figure 1: Graph of features of a personalized product.
Based on the constructed graph of the organization of the personalized product, its
morphological map can be developed. In the process of morphological analysis, the artificial
intelligence system can select features or elements. Fig. 2 shows a diagram of the algorithm for
creating a personalized product with the following components. A dress (С0) for everyday use
(С1,11) for women aged 25-30 (С1,23) for wearing in the spring-autumn period (С1,33). The dress
must have patch pockets (С1,61), a central button fastening (С1,62) and a hood (С1,63). The style of
the dress is classic (С2,11), the color of the main material (С2,2) is blue.
Individual parts of the product are decorated with sequin appliqué (С2,43,2) and rhinestone
appliqué (С2,43,3). The silhouette is fitted (С2,51), slightly widened to the bottom. The sleeve is sewn
in (С2,61), elbow-length, with a wide cuff at the bottom of the sleeve. Dress for a business class
woman (С3,12) made of natural materials (С4,11), which ensures ecological production (С3,41). Matte
fabric (С4,24). The product includes a sensor for monitoring physical parameters of a person (С4,43).
The price range of the dress is average (С5,12). A product of a well-known brand (С5,21).
Figure 2: General morphological map of creating a personalized product.
Based on the morphological map, such a product can be defined as:
W = [C0, (С1,11˄С1,23˄С1,33˄С1,61˄С1,62˄С1,63), (С2,11˄С2,2˄С2,43,2˄С2,43,3˄С2,51˄С2,61), (С3,12˄
˄С3,41), (С4,11˄С4,24˄С4,43), (С5,12˄С5,21)].
An example of creating a personalized product based on the algorithm is shown in Figure 3.
Figure 3: A personalized product created using artificial intelligence based on a developed
algorithm.
Considering the number of finite elements in each group, it is possible to predict a fairly large
number of combinations of the main features and the appearance of personalized products with
new properties. As a preliminary step, it is necessary to determine the basic rules according to
which individual characteristics can be combined in a separate group. Obviously, the elements
of each group cannot be chosen arbitrarily. Examples of possible combinations for the first
group of features are given in Table 1.
Table 1
Combinations of group C1 are possible
С1,11 С1,12 С1,21,1 С1,21,1 С1,22,1 С1,22,2 С1,31 С1,32 С1,61 С1,62
С1,11 ˅ ˅ ˅ ˄ ˄ ˅ ˅ ˄˅ ˄˅
С1,12 ˅ ˅ ˅ ˄ ˄ ˅ ˅ ˄˅ ˄˅
С1,21,1 ˅ ˅ ¬ ˄ ˄ ˅ ˅ ˄˅ ˄˅
˅ ˅ ˄ ˄ ˅ ˅ ˄˅ ˄˅
С1,21,2 ¬
˄ ˄ ˄
¬ ˅ ˅ ˄˅ ˄˅
С1,22,1 ˄
˄ ˄ ˄ ¬ ˅ ˅ ˄˅ ˄˅
С1,22,2 ˄
˅ ˅ ˅ ˅ ˅ ˄˅ ˄˅
С1,31 ˅ ˄˅ ˄˅
˅ ˅ ˅ ˅ ˅ ˅
С1,32 ˅ ˄˅
˄˅ ˄˅ ˄˅ ˅ ˄˅ ˄˅ ˄˅
˄˅
С1,61 ˄˅ ˄˅ ˄˅ ˄˅ ˄˅ ˄˅ ˄˅ ˄˅
˄˅
С1,62 ˄˅
General approaches to the formation of algorithms for the selection of a personalized product
will be based on the calculation of statements with the help of sentential relations. Complex
systems of personalized products consist of elements of the graph of morphological features. In
a formalized form, each product will represent a complex statement consisting of simple signs
connected by means of actions that can be expressed sententially:
- The conjunction ("and") is used to indicate the simultaneous presence of several
features. For example, "the presence of pockets and a hood" indicates that both elements are
present in the product;
- Disjunction ("or") is used to choose between alternatives. For example, "sewn sleeve or
raglan sleeve" means that the product can have one of these two types of sleeves;
- The negation ("not") is used to indicate the absence of certain features. For example, "no
collar" indicates that the product does not have this element;
- Implication ("if... then") - this relationship allows you to determine the dependence
between features. For example, "if the product does not have an insulating lining, then it is
intended for summer" indicates that the absence of an insulating lining is an indicator of the
product's seasonality.
These words or combinations of words are sentential connections for which the following
symbols are traditionally used:
¬ - for "no", ˄ - for "and", ˅ - for “or”, → - for "if...then".
The mandatory list of the main features characterizing a fashion product as a structurally
homogeneous object looks like this: (С1˄С2˄С3˄С4˄С5).
The use of simplified formulas and approaches to communication with the artificial
intelligence system leads to unsatisfactory results. A possible result of the application of
simplified approaches is illustrated in Fig. 4, which, of course, lags far behind the results
presented in Fig. 3, which were obtained on the basis of the developed graph.
Figure 4: Unsatisfactory appearance of the model synthesized using simplified formulas.
This indicates that in order to achieve more accurate and effective results in interaction with
artificial intelligence systems, it is necessary to use more complex and formalized approaches.
In order to formalize the described properties of combinations of the main features, we will
present them in mathematical form. Then the main subgroups of features for the first group
based on Figs. 2 will be presented in the form:
С1,11˅С1,12; С1,11˅С1,22,1; (С1,11˄С1,61) (С1,11˅С1,61);
С1,21,1 С1,12; С1,21,1 ¬ С1,21,2; С1,211˅С1,31; (С1,21,2˄С1,61) (С1,21,2˅С1,61)
Similar expressions can be written for all groups. By combining features and formulas, it is
possible to synthesize a variety of personalized products with various combinations of features
selected from the list of rational combinations.
3. Conclusions
To ensure the functioning of artificial intelligence systems for creating personalized products, a
morphological analysis of potential products, presented as a feature system, was carried out.
Five groups, including 114 features, were identified. Basic rules and formal approaches to
combining these key features were established. A graph of personalized product features was
developed, along with the rationale for the main connections, which formed the basis for
creating a morphological data table used for communication with the AI system. An analysis of
the depth of use of this morphological data in the AI system was conducted. The effectiveness of
using logical connections in forming algorithms for selecting personalized products was
demonstrated, allowing for the creation of flexible and adaptive models. These models can
consider various combinations of features, improving the accuracy of forecasting consumer
needs and enhancing the effectiveness of personalization in the fashion industry.
In contrast to known advancements in artificial intelligence, the proposed model
incorporates a complex structure that includes precise indicators for clothing characteristics.
Machine learning algorithms, including spectral graph filtering and recurrent neural networks,
were used to train the model. This enabled the system to analyze data and identify relationships
between clothing characteristics, facilitating the creation of personalized recommendations for
users.
The model is capable of self-improvement during operation through machine learning on
new data sets, allowing for the automatic addition of new characteristics and parameters to the
existing graph structure.
References
[1] R. D. Baruah, M. M. Organero, Explicit Context Integrated Recurrent Neural Network for
applications in smart environments, Expert Systems with Applications (2024) volume 255.
doi.org/10.1016/j.eswa.2024.124752.
[2] A. Khalilipour, F. Bozyigit, M. Challenger, "Intelligent Model Management based on
Textual and Structural Extraction-An Exploratory Study," 2024 10th International
Conference on Web Research (ICWR) pp. 165-174.
doi.org/10.1109/ICWR61162.2024.10533323.
[3] P. Jing, K. Cui, J. Zhang, Y. Li, Y. Su, "Multimodal High-Order Relationship Inference
Network for Fashion Compatibility Modeling in Internet of Multimedia Things," in IEEE
Internet of Things Journal 01.01 (2024) pp. 353-365.
doi.org/10.1109/JIOT.2023.3285601.
[4] H. A. D. M. Arachchi, G. D. Samarasinghe, Intention to Adopt Intelligent Clothing in the
Fashion Retail Industry: Extending the HISAM Model with Technology Readiness,
International Journal of Human–Computer Interaction (2023) pp. 1–15.
doi.org/10.1080/10447318.2023.2254622.
[5] M. Riabchykov, V. Mytsa, O. Tkachuk, O. Pakholiuk, D. Melnyk, "Efficiency of Protective
Textile Smart Systems Using Electronic Tags," in: Nechyporuk, M., Pavlikov, V., Krytskyi,
D. (eds), Integrated Computer Technologies in Mechanical Engineering - 2023. ICTM 2023.
Lecture Notes in Networks and Systems, volume 1008, Springer, Cham.
doi.org/10.1007/978-3-031-61415-6_16.
[6] M. Riabchykov, A. Alexandrov, R. Trishch, A. Nikulina, N. Korolyova, Prospects for the
Development of Smart Clothing with the Use of Textile Materials with Magnetic
Properties, Tekstilec (2022) pp. 36-43. doi.org/10.14502/tekstilec.65.2021050.
[7] A. Jiju, A. Anilkumar, K. P. Gokul Krishnan, J. George and R. S. Prasanth, "Review on
Fashion Trend Analysis and Forecasting Techniques - A Machine Learning Approach,"
2024 5th International Conference on Innovative Trends in Information Technology
(ICITIIT) pp. 1-6. doi.org/10.1109/ICITIIT61487.2024.10580247.
[8] J. Huang, X. Zhang, Application of fashion element trend prediction model integrating am
and efficientnet-b7 models in art design, Scalable Computing: Practice and Experience
(2024) pp. 2756–2769. doi.org/10.12694/scpe.v25i4.2814.
[9] A. Bompotas, P. Triantafyllopoulos, G. E. Raptis, C. Katsini, C. Makris, "Towards exploring
personalized hyperlink recommendations through machine learning," UMAP 2024 -
Adjunct Proceedings of the 32nd ACM Conference on User Modeling, Adaptation and
Personalization, pp. 528–533. doi.org/10.1145/3631700.3664913.
[10] D.-N. Nguyen, V.-H. Nguyen, T. Trinh, T. Ho, H.-S. Le, A personalized product
recommendation model in e-commerce based on retrieval strategy, Journal of Open
Innovation: Technology, Market, and Complexity (2024) volume 10, 100303.
doi.org/10.1016/j.joitmc.2024.100303.
[11] X. Cong, W. Zhang, The application of hierarchical perception technology based on deep
learning in 3D fashion design, Heliyon (2024) e29983.
doi.org/10.1016/j.heliyon.2024.e29983.
[12] A. Vuruskan, G. Demi̇Rki̇Ran, E. Bulgun, T. Ince, C. Guzelis, Design of an interactive
fashion recommendation platform with intelligent systems, Proiectarea unei platforme
interactive de recomandare a articolelor de modă cu sisteme inteligente, Industria Textila
(2024) pp. 177–184. doi.org/10.35530/IT.075.02.202312.
[13] M. Riabchykov, V. Mytsa, M. Bondarenko, T. Popova, S. Nechipor A. Nikulina, S.
Bondarenko, Formation of complex 3D surfaces scans for garment CAD, Vlakna a Textil
(2023) pp. 13-18. doi.org/10.15240/tul/008/2023-3-002.
[14] M. Riabchykov, S. Vilkov, S. Nechipor, T. Popova, Two-dimensional studies of
thermomechanical properties of textile materials for 3D formation, Vlakna a Textil (2018)
pp. 87-92. http://vat.ft.tul.cz/Archive/VaT_2018_2.html.
[15] A. Jiju, A. Anilkumar, K. P. Gokul Krishnan, J. George, R. S. Prasanth, "Review on Fashion
Trend Analysis and Forecasting Techniques - A Machine Learning Approach," 2024 5th
International Conference on Innovative Trends in Information Technology (ICITIIT) pp. 1-
6. doi.org/10.1109/ICITIIT61487.2024.10580247.
[16] G. Skenderi, C. Joppi, M. Denitto, M. Cristani, Well googled is half done: Multimodal
forecasting of new fashion product sales with image-based google trends, Journal of
Forecasting (2024). doi.org/10.1002/for.3104.
[17] F. Rooholamini, R. Azmi, M. Khademhossein, M. Zarvani, "Dynamic Knowledge Enhanced
Neural Fashion Trend Forecasting with Quantile Loss," 2023 13th International Conference
on Computer and Knowledge Engineering (ICCKE) pp. 601606.
doi.org/10.1109/ICCKE60553.2023.10326316.
[18] N. Thoiba Singh, L. Chhikara, P. Raj, S. Kumar, "Fashion Forecasting using Machine
Learning Techniques," 2023 IEEE International Conference on Integrated Circuits and
Communication Systems (ICICACS) pp. 01-06.
doi.org/10.1109/ICICACS57338.2023.10099844.
[19] C. Muttaraju, R. N. Prabhu, S. Sheetal, D. Uma, S. S. Shylaja, "IFF: An Intelligent Fashion
Forecasting System," in: Buyya, R., Hernandez, S. M., Kovvur, R. M. R., Sarma, T.H. (eds)
Computational Intelligence and Data Analytics. Lecture Notes on Data Engineering and
Communications Technologies (2023) volume 142, Springer, Singapore.
doi.org/10.1007/978-981-19-3391-2_37.
[20] S-I. Papadopoulos, C. Koutlis, S. Papadopoulos, I. Kompatsiaris, Multimodal
QuasiAutoRegression: forecasting the visual popularity of new fashion products,
International Journal of Multimedia Information Retrieval (2022) 11, pp. 717–729.
doi.org/10.1007/s13735-022-00262-5.
[21] M Riabchykov, V. Mytsa, K. Ryabchykova, "Artificial Intelligence as a Tool for the
Development of Professional Competencies of a Fashion Industry Specialist," in: Faure, E.,
et al. Information Technology for Education, Science, and Technics. ITEST 2024. Lecture
Notes on Data Engineering and Communications Technologies (2024) volume 222,
Springer, Cham. https://doi.org/10.1007/978-3-031-71804-5_20.
[22] D. Mo, X. Zou, WK. Wong, Neural stylist: Towards online styling service, Expert Systems
with Applications (2022) volume 203, 117333. doi.org/10.1016/j.eswa.2022.117333.
[23] Y. Ma, Y. Ding, X. Yang, L. Liao, WK. Wong, T.-S. Chua, “Knowledge enhanced neural
fashion trend forecasting,” ICMR 2020 - Proceedings of the 2020 International Conference
on Multimedia Retrieval, pp. 82–90. doi.org/10.1145/3372278.3390677.
[24] E, Vishal, E. Ushaa, E. Vivek, M. Keerthna, Revolutionizing Healthcare: The Application of
Image Processing Techniques (2024). 10.4018/979-8-3693-2105-8.ch019.
[25] X. Liu, Y. Hao, L. Zhao, G. Liu, V. S. Sheng, P. Zhao, LMACL: Improving Graph
Collaborative Filtering with Learnable Model Augmentation Contrastive Learning, ACM
Transactions on Knowledge Discovery from Data (2024) 18(7), 177.
doi.org/10.1145/3657302.
[26] O. Alshareet, A. Ben Hamza, Adaptive spectral graph wavelets for collaborative filtering.
Pattern Analysis and Applications (2024). doi.org/10.1007/s10044-024-01214-x.
[27] E. De Santis, A. Rizzi, Modeling failures in smart grids by a bilinear logistic regression
approach, Neural Networks (2024) volume 174, 106245.
doi.org/10.1016/j.neunet.2024.106245.
[28] H. Cengiz, M. Şenel, The effect of perceived scarcity on impulse-buying tendencies in a fast
fashion context: A mediating and multigroup analysis, Journal of Fashion Marketing and
Management (2024) 28(3), рр. 405–425. doi.org/10.1108/JFMM-03-2023-0082.
[29] A. Slavinska., S. Matiukh, V. Mytsa, O. Syrotenko, O. Dombrovska, Devising a method for
the interactive arrangement of structural elements of men’s jacket models,
EasternEuropean Journal of Enterprise Technologies (2022) 1, 1 (115), рр. 56–66.
doi.org/10.15587/1729-4061.2022.251065.
[30] Y. K. Lee, How complex systems get engaged in fashion design creation: Using artificial
intelligence. Thinking Skills and Creativity (2022), 46.
https://doi.org/10.1016/j.tsc.2022.101137.
[31] S. Karadayi-Usta, Role of artificial intelligence and augmented reality in fashion industry
from consumer perspective: Sustainability through waste and return mitigation,
Engineering Applications of Artificial Intelligence (2024) 108114.
doi.org/10.1016/j.engappai.2024.108114.
[32] J. Hu, F. El-Rashid, S. E. Bertelsen, “SHADE: Empowering Consumer Choice for Sustainable
Fashion with AI and Digital Tooling,” CHI EA '24: Extended Abstracts of the 2024 CHI
Conference on Human Factors in Computing Systems, 282.
doi.org/10.1145/3613905.3651025.