Production Demand Forecasting in the Textile Industry

Using Machine Learning Techniques

A thesis submitted in partial fulfilment of the requirements for

the award of the degree of

B.Tech
in
Production Engineering

by
Arunprabhakar P (114120020)
Iyyappan BG (114120050)
Varun M (114120076)

PRODUCTION ENGINEERING
NATIONAL INSTITUTE OF TECHNOLOGY
TIRUCHIRAPPALLI – 620015

MAY 2024

i
 BONAFIDE CERTIFICATE

This is to certify that PRIR19 - PROJECT WORK the project titled

“Production Demand Forecasting in the Textile Industry using Machine
Learning Techniques” is a bonafide record of the work done by

Arunprabhakar P (114120020)

Iyyappan BG (114120050)

Varun M (114120076)

in partial fulfilment of the requirements for the award of the degree of Bachelor of Technology
in Production Engineering of the NATIONAL INSTITUTE OF TECHNOLOGY,
TIRUCHIRAPPALLI, during the year 2020-2024.

Project Guide Project Co-guide Head of the Department

Project Viva-voce held on_____________________

Internal Examiner External Examiner

ii
 ABSTRACT
The textile industry currently analyses historical sales data from individual stores (e.g., 3,000

stores) to forecast demand for upcoming seasons. While leveraging historical trends, this

approach is time-consuming and labour-intensive and needs help to account for broader market

shifts and customer preferences. This project proposes a new method for production demand

forecasting that leverages machine learning and incorporates internal and external factors.

We explore the limitations of using isolated historical sales data and propose a machine

learning-based approach that considers internal factors like production lead time and lost time

alongside external factors like GDP and customer buying ability. By collaborating with textile

vendors, we identified relevant data points and developed machine learning models, including

linear regression and SARIMA. Our results demonstrate that incorporating these additional

factors improves the accuracy of demand forecasts compared to traditional methods.

This project offers a more comprehensive and efficient approach to production demand

forecasting in the textile industry. By leveraging machine learning and a more extensive range

of data points, companies can better understand customer buying ability and optimise

production processes, reducing inventory costs and increasing competitiveness.

Keywords: Textile Industry, Production Demand Forecasting, Machine Learning Techniques,

SARIMA

iii
 Acknowledgement

We would like to express our sincere gratitude to all those who have contributed to our research

on production demand forecasting in the textile industry using machine learning techniques.

First and foremost, we extend our heartfelt thanks to our supervisor for their guidance, support,

and invaluable insights throughout the project. Their expertise has been instrumental in shaping

our research methodology and analysis. We would also like to thank Rycla Overseas for

providing essential data and insights into the textile industry's operations. Their contribution

has been crucial in understanding the intricacies of demand forecasting within the industry.

We express our gratitude to Professors Dr. Matruprasad Rout and Dr.Parthiban for being

our Guide and Co-Guide respectively. We sincerely thank scholar Kamlesh Pant for assisting

us in the completion of the project.

Additionally, we express our appreciation to our colleagues for their helpful suggestions and

feedback, which have significantly enhanced the quality of our paper. Their insights have

played a vital role in refining our machine-learning techniques and improving the accuracy of

our forecasting models. To all who have contributed, your support has been invaluable, and we

are truly grateful for your assistance in this endeavour.

iv
 TABLE OF CONTENTS

Title Page No.

ABSTRACT iii

ACKNOWLEDGEMENT iv

TABLE OF CONTENTS v-vi

LIST OF FIGURES vii

LIST OF TABLES viii

ABBREVIATIONS ix

CHAPTER 1 INTRODUCTION

1.1 The Ever-Shifting Landscape of the Textile Industry 1

1.2 Limitations of Traditional Forecasting Methods 1

1.3 The Need for a More Comprehensive Approach 1

1.4 Objectives 1-2

CHAPTER 2 METHODOLOGY

2.1. Data Collection 3

2.2. Exogenous and Endogenous Factor Identification 3-4

2.3. Forecasting Model Selection 5

2.4. Model Development 5

2.5. Error Analysis and Validation 6

2.6. Forecast Generation 6

2.7. Sensitivity Analysis 6

CHAPTER 3 LITERATURE REVIEW

3.1. Limitations of Traditional Forecasting Methods 7

3.2 Machine Learning for Demand Forecasting 8

3.3 Existing Research on ML in Textile Demand Forecasting 8

3.4 Gaps in Existing Literature 9

v
3.5 Contribution of this Research 9

3.6. Industry Research 10

3.7. Approached Companies 10

3.8. Different types of textile industries 11

3.9. Key Insights on Demand 11-12

3.10 Data Acquisition 12

3.11. Moving Forward 13

3.12 Data and Authorisation Letter 13

CHAPTER 4 RESULTS AND DISCUSSION

4.1 Final Data Set 14

4.2 Model Selection and Development 15-16

4.3 Model Development Process 16-17

4.4 Machine Learning Model Codes and Results 17-28

4.5 Comparison of R-squared Values 29

CHAPTER 5 CONCLUSIONS 30

REFERENCES 31-32

vi
 LIST OF FIGURES

Figure No. Title Page No.

1 Research Methodology 4

2 Project Timeline 6

3 Final Data Set 14

4 Production in 2023 14

5 Linear Regression Model considering without Factors 17

6 Linear Regression Model considering with factors 18

7 ARIMA Model considering with factors 19

8 Code for SARIMA Model considering without Factors 20

9 Output for SARIMA Model considering without Factors 21

10 Code for SARIMA Model considering with Factors 22

11 Output for SARIMA Model considering with Factors 23

12 Code for Polynomial Regression Model considering without factors 24

13 Output for Polynomial Regression Model considering without factors 25

14 Code for Polynomial Regression Model considering with factors 26

15 Output for Polynomial Regression Model considering with factors 27

16 Random Forest Model considering with Factors 28

17 Comparison of Forecasted Production for different Models 29

vii
 LIST OF TABLES

Table No. Title Page No.

1 Comparison of R Square Values 29

viii
 ABBREVIATIONS

SARIMA Seasonal Auto Regressive Integrated Moving Average

SD Standard Deviation
LR Linear Regression
ML Machine Learning
MSE Mean Squared Error
RMSE Root Mean Squared Error
KNN K Nearest Neighbours
GDP Gross Domestic Product
GWP Global Warming Potential
R Square Coefficient of Determination

ix
 CHAPTER 1

INTRODUCTION

1.1 The Ever-Shifting Landscape of the Textile Industry

The textile industry has historically navigated a dynamic landscape, constantly adapting to
meet consumers' ever-evolving needs and desires. Today, this dynamism is amplified by the
surge of e-commerce, where fashion trends can change rapidly and unpredictably. In this fast-
paced environment, accurately forecasting production demand has become a critical
differentiator for textile manufacturers. Traditional methods rely on historical sales data from
individual stores and must be revised.

1.2 Limitations of Traditional Forecasting Methods

While offering a baseline for production planning, traditional forecasting methods have several
limitations. Firstly, they are time-consuming and labour-intensive, requiring significant manual
effort to analyse vast data. Secondly, these methods focus solely on historical sales data, failing
to capture the broader market shifts and customer preferences that drive demand fluctuations.
They neglect external factors like economic conditions, social media trends, and competitor
activity, which can significantly impact consumer buying behaviour.

1.3 The Need for a More Comprehensive Approach

Manufacturers require a more comprehensive and efficient approach to production demand

forecasting to thrive in the current textile industry landscape. This project proposes a novel
solution that leverages the power of machine learning (ML) techniques. By incorporating both
internal (endogenous) and external (exogenous) factors that influence production demand, our
project aims to develop a more accurate and efficient forecasting model.

1.4 Objectives

Our main objective in this work is to develop and evaluate demand forecasting models tailored
to specific textile products. By carefully comparing the performance of these models, we aim
to identify the most effective demand forecasting methods in the textile industry. We will
achieve this by:

1
● Developing ML-based forecasting models: We will explore and implement various
machine learning models suitable for time series forecasting, incorporating both
internal and external factors.
● Evaluating model performance: We will assess the accuracy and effectiveness of
each model through rigorous error analysis and validation techniques.
● Identifying impactful factors: We will evaluate how endogenous and exogenous
factors influence production demand, providing valuable insights for optimising
production processes.
● Providing actionable insights: Our ultimate goal is to equip apparel manufacturers
with actionable insights that will enable them to:
○ Optimise manufacturing processes to meet actual demand.
○ Reduce excess inventory, leading to cost savings.
○ Improve their competitiveness in the market by anticipating trends and adapting
production accordingly.
○ Increase market share by effectively catering to customer needs.

2
 CHAPTER 2

METHODOLOGY

This project will employ a data-driven approach to develop and evaluate machine learning

models for production demand forecasting in the textile industry. The methodology will be

divided into the following key stages:

2.1. Data Collection

● Internal Data Acquisition: We will collaborate with textile vendors to gather

production data relevant to our analysis. This data may include:

○ Historical production quantities for specific textile products (e.g., T-shirts)

○ Internal factors that can influence production lead times, such as:

■ Departmental flow times

■ Lost time due to machine downtime or other disruptions

■ Process times for different manufacturing stages

● External Data Acquisition: We will collect data on external factors that may influence
production demand. Sources for this data may include:

○ Government websites and economic reports

○ Financial databases

○ Social media listening tools (to gauge consumer trends)

○ Industry publications

○ Publicly available datasets

2.2. Exogenous and Endogenous Factor Identification

● Literature Review: We will conduct a comprehensive review of existing research to

identify factors that have been demonstrated to influence production demand in the

textile industry. This review will encompass both endogenous and exogenous factors.

3
● Industry Expert Consultation: We will consult with textile industry experts,

including manufacturers and analysts, to gain insights into the most relevant factors

affecting production demand in their specific areas of expertise.

● Data-Driven Analysis: We will explore the relationships between the collected data

points (production quantities, internal factors, external factors) using statistical

techniques. This will help us identify the factors with the strongest correlations

production demand.

Linear Regression

Random Forest

SARIMA

Polynomial Regression

Figure 1 - Research Methodology

4
2.3 Forecasting Model Selection

Based on the characteristics of our data and the research conducted, we will select appropriate

machine learning models for demand forecasting. Here are some potential models to consider:

● Linear Regression: A foundational model that establishes a linear relationship

between production quantities and a set of independent variables (internal and external

factors).

● Time Series Models: These models analyse historical data patterns to forecast future

trends. Examples include ARIMA (Autoregressive Integrated Moving Average) and

SARIMA (Seasonal ARIMA).

● Other Machine Learning Models: Depending on the complexity of the data and the

desired level of accuracy, we may explore more advanced models such as Random

Forests or Polynomial Regression.

2.4. Model Development

● Data Preprocessing: The collected data will be carefully pre-processed to ensure its

quality and suitability for machine learning model development. This may involve

handling missing values, identifying and addressing outliers, and potentially scaling

features for better model performance.

● Model Training and Testing: The chosen machine learning models will be trained
and tested on a split of the data. The training data will be used to fit the model

parameters, while the testing data will be used to evaluate the model's generalizability

and ability to forecast unseen data.

● Hyperparameter Tuning: We will optimise each model's hyperparameters (adjustable

settings) to achieve the best possible performance. This may involve techniques like

grid search or randomised search.

5
2.5. Error Analysis and Validation

● Evaluation Metrics: We will assess the performance of each model using appropriate

error metrics like Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and

R-squared. These metrics will quantify the difference between the model's forecasts

and production values.

● Model Comparison: We will compare the performance of different models on the

testing data to identify the model with the highest accuracy and lowest error rate.

● Validation: To ensure the generalizability of the chosen model, we may employ

additional validation techniques such as k-fold cross-validation.

2.6. Forecast Generation

Once the best-performing model is selected and validated, we will use it to generate future

production demand forecasts. These forecasts will be based on the most recent production data

and the current values of the identified exogenous factors.

2.7. Sensitivity Analysis

We may conduct a sensitivity analysis to understand how changes in individual factors (both

internal and external) impact the model's forecasts. This can provide valuable insights for

production planning and resource allocation.

Figure 2 - Timeline of Project with Gantt Chart Representation

6
 CHAPTER 3
LITERATURE REVIEW

The textile industry thrives on its ability to anticipate and cater to ever-evolving customer

demands. However, in the fast-paced world of e-commerce, where fashion trends shift rapidly,

accurate production demand forecasting has become critical for effective inventory

management and successful business operations. Traditional methods, which rely solely on

historical sales data from individual stores, are becoming increasingly insufficient. This

literature review explores the limitations of traditional methods and highlights the potential of

machine learning techniques for more comprehensive and efficient production demand

forecasting in the textile industry.

3.1 Limitations of Traditional Forecasting Methods

Traditional forecasting methods, while offering a baseline for production planning, suffer from

several limitations:

● Time-consuming and Labor-intensive: These methods typically involve manual

analysis of vast amounts of data, requiring significant time and resources.

● Limited Scope: They focus solely on historical sales data, neglecting broader market

trends and customer preferences driven by social media, economic conditions, and

competitor activity.

● Inaccuracy for Non-Seasonal Products: Traditional methods struggle to adapt to

rapidly changing trends and may not accurately forecast demand for non-seasonal or

quickly evolving fashion styles.

7
3.2 Machine Learning for Demand Forecasting

Machine learning (ML) offers a promising approach to overcoming the limitations of

traditional methods. ML algorithms can learn from historical data and identify complex

relationships between various factors that influence production demand. This enables them to

generate more accurate and flexible forecasts. Here's how ML is being utilized in demand

forecasting:

● Improved Accuracy: Studies have shown that ML models can outperform traditional

methods in terms of forecasting accuracy

● Incorporation of Multiple Factors: ML models can integrate both internal

(production lead time, lost time) and external factors (economic indicators, social

media trends) for a more holistic view of demand drivers.

● Adaptability: ML models can adapt to changing trends and market dynamics by

continuously learning from new data

3.3 Existing Research on ML in Textile Demand Forecasting

Several studies have explored the application of machine learning for production demand

forecasting in the textile industry:

● Lorente-Leyva et al. (2021, 2020): These studies employed machine learning models

that incorporate both endogenous (internal) and exogenous (external) factors for

textile demand forecasting, demonstrating the effectiveness of this approach.

● Xin et al. (2021): This research investigated the use of machine learning methods like

K-Nearest Neighbours (KNN) for forecasting demand for textile products, indicating

the potential of various ML algorithms in this domain.

8
3.4 Gaps in Existing Literature

While existing research demonstrates the potential of ML in textile demand forecasting, some

gaps remain:

● Limited Focus on Specific Products: Most studies focus on general textile demand
forecasting, neglecting the need for tailored models for specific product categories

(e.g., T-shirts vs. Formal wear).

● Evaluation of Impactful Factors: While research acknowledges the importance of

internal and external factors, a deeper understanding of which factors have the most

significant influence on demand for different textile products is still needed.

3.5 Contribution of this Research

This project aims to address these gaps by:

● Developing Tailored ML Models: We will develop and evaluate ML models

specifically designed for forecasting demand for different textile products.

● Identifying Impactful Factors: We will analyse the impact of various internal and

external factors on production demand for specific textile product categories.

● Providing Actionable Insights: Our aim is to provide apparel manufacturers with

actionable insights to optimize production processes, reduce inventory costs, and

improve market competitiveness by effectively meeting customer demand.

By addressing these limitations and contributing new knowledge, this research seeks to

advance the application of machine learning for production demand forecasting in the textile

industry.

9
3.6 Industry Research

Challenges and Collaborations

Our project necessitates a comprehensive understanding of the textile industry's production and
supply chain dynamics. To achieve this, we initiated contact with five textile vendors. Our
primary objectives were twofold:

1. Gaining Business Model Insights: We sought to delve into the vendors' business
models, specifically exploring their product development processes and distribution
channels.
2. Data Acquisition for Forecasting: Accessing data relevant to production demand
forecasting was crucial for developing our machine learning models.

While three of the vendors (Asna Garments, Ammaiyappar Textiles, and Sri Geetha Textiles)
declined to participate, valuable information and collaboration were established with
Bhuvaneshwari Textiles and Rycla Overseas. Their willingness to share insights provided a
deeper understanding of the industry's challenges in data collection for demand forecasting
purposes.

3.7 Approached Companies

To illustrate the complexities and best practices within the textile industry, we detail our
interactions with the following companies:

1. Asna Garments, Tiruppur

2. Ammaiyappar Textiles, Tiruvallur
3. Bhuvaneshwari Textiles, Dindugal
4. Sri Geetha Textiles, Hosur
5. Rycla Overseas, Tiruppur

10
3.8 Different types of textile industries:
● Bhuvaneshwari Textiles (Lungis): Their meticulously planned lungi production
process exemplifies their commitment to both quality and efficiency. Despite a fixed
schedule, they consistently maintain predictable monthly production volumes without
any major variation in their monthly production. This interaction provides valuable
insights into a well-managed production environment, highlighting the importance of
standardized processes for consistent output.
Monthly Production: 400 per Power loom Machine

● Rycla Overseas (T-shirts, Track Pants, etc.): Their diverse product portfolio and
robust export network demonstrate their ability to cater to a wide range of consumer
needs across different demographics and geographies. This interaction showcases a
multifaceted approach to production and sales, highlighting the adaptability and market
focus required for success in the global textile industry.

● Kabooki (Textile Store, Denmark): Their collaboration with Rycla Overseas

showcases a successful partnership between a manufacturer and an importer.
Kabooki's meticulous design, forecasting, and quality-focused approach highlight
their commitment to delivering high-quality apparel. This interaction emphasizes the
importance of planning, collaboration, and a focus on customer needs throughout the
supply chain.

3.9 Key Insights on Demand

Through our interactions with these companies, we were able to glean valuable insights into
the factors that significantly affect demand in the textile industry:

● Price and Quality: The example of leggings (where Twin Birds succeeded with a
quality focus and Prisma captured the market with lower pricing) underscores the
importance of considering both price and quality in our demand forecasting models.
Consumers often make purchasing decisions based on a balance of these two factors.

11
 ● Non-Productive Time: Garment production facilities often track lost time
(downtime) to optimise production efficiency. This metric, while not readily available
in this instance, can be a valuable input for production planning, as it reflects potential
bottlenecks or areas for improvement within the manufacturing process.

3.10 Data Acquisition

● Endogenous Factors: Our efforts to acquire data on internal factors like production
lead time and lost time proved challenging. The vendors did not have readily available
or well-recorded data on these factors, highlighting a potential gap in their data
collection practices. We will need to explore alternative methods for obtaining this type
of data, such as conducting surveys or on-site observations within textile manufacturing
facilities.

● Exogenous Factors: Data on external factors like GDP and Exchange rates was
successfully obtained through web research. These factors can significantly impact
production demand by influencing consumer spending power and import/export costs.

ENDOGENOUS FACTORS:

● Production Lead Time

● Departmental Flows
● Lost Time
● Process Times

EXOGENOUS FACTORS:

● Exchange Rates
● Interest Rates
● GDP
● GWP
● Repo Rate

12
3.11 Moving Forward

Despite the challenges in obtaining comprehensive endogenous data, the insights from
collaborating with the vendors and the collected exogenous data will provide a solid foundation
for developing our machine learning models for production demand forecasting. We will
continue to explore alternative methods for acquiring internal data points and refine our data
collection strategy to ensure the most accurate and comprehensive forecasting models possible.

3.12 Data and Authorisation Letter

13
 CHAPTER 4

RESULTS AND DISCUSSION

4.1 Final Data Set:

Figure 3 – Final Data Set

We focus only on the Production of T Shirts category in our analysis. In the final data set, the
‘PRODUCTION” field refers to the total number of T Shirts produced per month. Rest are self-
explanatory Exogenous Factors.

Figure 4 – Production in 2023

14
4.2 Model Selection and Development:

Our project will explore several machine learning models for production demand forecasting
in the textile industry. We will compare their performance to identify the most effective
approach for our specific data and forecasting needs. Here's a breakdown of the models we will
consider:

1. Linear Regression:

● Model Description: Linear regression establishes a linear relationship between a

dependent variable (production quantity) and one or more independent variables
(internal and external factors). It's a good starting point for understanding the basic
relationships between production and influencing factors.
`
● Implementation in our project:

○ We will develop a linear regression model with production quantity as the

dependent variable.
○ We will initially build a model using only internal factors (e.g., production
lead time, lost time).
○ We will then expand the model to include external factors (e.g., GDP,
exchange rates) and compare the results.
○ The R-squared value will be used to assess the model's ability to explain the
variance in production quantity.

2. SARIMA (Seasonal Autoregressive Integrated Moving Average):

● Model Description: This time series model is well-suited for forecasting data with
seasonal patterns. It takes into account past values of the dependent variable
(production quantity), lagged error terms (differences between forecasts and actual
values), and seasonal components.

15
 ● Implementation in our project:

○ We will develop a SARIMA model using historical production data.

○ We will identify and incorporate the appropriate seasonal component (e.g.,
monthly, quarterly) based on the data's pattern.
○ We will evaluate the model's performance with and without including
exogenous factors using metrics like Mean Squared Error (MSE) and Root
Mean Squared Error (RMSE).

3. Additional Machine Learning Models:

Depending on the complexity of our data and the desired level of accuracy, we may explore
more advanced models:

● Random Forests: This ensemble learning technique combines multiple decision trees
to improve prediction accuracy and reduce overfitting. It can handle both numerical
and categorical variables.
● Gradient Boosting Machines: This sequential learning approach builds an ensemble of
decision trees, where each subsequent tree learns from the errors of the previous ones,
potentially leading to higher accuracy.

4.3 Model Development Process:

For each model chosen:

1. Data Preprocessing: Clean and prepare the data by handling missing values,
identifying outliers, and potentially scaling features for better model performance.
2. Model Training and Testing: Split the data into training and testing sets. Train the
model on the training data and evaluate its performance on the unseen testing data.
3. Hyperparameter Tuning: Optimize the model's hyperparameters (adjustable settings)
to achieve the best possible performance. Techniques like grid search or randomized
search can be employed.
4. Model Comparison: Evaluate and compare the performance of different models using
error metrics like MSE, RMSE, and R-squared. The model with the highest accuracy
and lowest error rate will be considered the best performing model.

16
By following these steps and comparing the results of various models, we can identify the most
effective approach for forecasting production demand in the textile industry, considering both
internal and external factors.

4.4 Machine Learning Model Codes and Results:

1. Linear Regression Model:

a. Considering without Factors:

R Square Value = 0.17

Figure 5 – Linear Regression considering without Factors

17
b. Considering with Factors:

R square Value = 0.68

Figure 6 – Linear Regression considering with factors

18
2. ARIMA Model:
a. Considering With Factors:

R Square Value = 0.49

Figure 7 – ARIMA Model considering with factors

19
3. SARIMA Model:

a. Considering without Factors:

Figure 8 – Code for SARIMA Model considering without Factors

20
 R square Value = 0.45

Figure 9 – Output for SARIMA Model considering without Factors

21
b. Considering with Factors:

Figure 10 - Code for SARIMA Model considering with Factors

22
 R Square Value = 0.998

Figure 11 – Output for SARIMA Model considering with Factors

23
4. Polynomial Regression Model:

a. Considering without Factors:

Figure 12 -- Code for Polynomial Regression Model considering without Factors

24
 R Square Value = 0.61

Figure 13 – Output for Polynomial Regression Model considering without factors

25
b. Considering with Factors:

Figure 14 -- Code for Polynomial Regression Model considering with Factors

26
 R Square Value = 0.63

Figure 15 – Output for Polynomial Regression Model considering with factors

27
5. Random Forest:
a. Considering with Factors:

R Square Value = 0.896

Figure 16 – Random Forest Model considering with Factors

28
 Figure 17 – Comparison of Forecasted Production for different Models

4.5 Comparison of R-squared Values:

MODEL/R SQUARE CONSIDERING CONSIDERING WITH

VALUE WITHOUT FACTORS FACTORS

LINEAR REGRESSION 0.17 0.68

ARIMA MODEL - 0.49

SARIMA MODEL 0.45 0.998

POLYNOMIAL 0.61 0.63

REGRESSION

RANDOM FOREST - 0.896

Table 1 – Comparison of R Square Values

Insight:

SARIMA model considering the factors has been found to be the best of all the models
considered.
29
 CHAPTER 5

CONCLUSIONS

We applied five machine learning models and assessed their accuracy. Firstly, we examined
Linear Regression and SARIMA models, both with and without considering the effect of
factors. Additionally, we developed a Polynomial Regression model, ARIMA and Random
Forest model.

Upon comparing the accuracy of all models, the SARIMA model, when factoring in the effects
of relevant factors, emerged as the most accurate among those considered.

In conclusion, our research highlights the importance of accurate demand forecasting for
achieving high-performance textile production. By mitigating production fluctuations, aligning
production with customer needs, and optimizing resource allocation (machinery, labor), a well-
performing forecasting model can significantly improve overall productivity. This translates to
a more stable, efficient, and customer-centric operation within the textile industry.

30
 REFERENCES

References

● [1] Desore, A., & Narula, S. A. (2018, August 21). Demand Forecasting for Textile
Products Using Machine Learning Methods.
https://www.researchgate.net/publication/220208887_Machine_Learning-
Based_Demand_Forecasting_in_Supply_Chains
● [2] Nenni, M. E., Giustiniano, L., & Pirolo, L. (2013, June). Demand Forecasting in
the Fashion Industry: A Review.
https://www.researchgate.net/publication/268386280_Demand_Forecasting_in_the_F
ashion_Industry_A_Review
● [3] Lorente-Leyva, J. A., Donate-Pérez, M. J., & Martínez-Alda, J. V. (2021).
Machine learning–assisted efficient demand forecasting using endogenous and
exogenous indicators for the textile industry.
https://www.tandfonline.com/doi/abs/10.1080/13675567.2022.2100334
1. Sabir, Emel Ceyhun & Esma, Batuk. (2013). Demand Forecasting Withof Using
Time Series Models In Textile Dyeing-Finishing Mills. Tekstil ve Konfeksiyon. 23.
143-151.
2. Textiles Intelligence (2020). Global Textile Industry Outlook to 2025.
https://indiantextilejournal.com/the-future-of-textiles-a-look-at-the-current-state-of-
affairs/#:~:text=%E2%A6%81%20The%20global%20apparel%20market%20is%20e
xpected%20to%20grow%20at,for%20comfortable%20and%20stylish%20clothing.
3. Aburto L and Weber L (2007). Improved Supply Chain Management Based on
Hybrid Demand Forecasts,Applied Soft Computing, Vol. 7, pp. 136-144.
4. Chawla, Aditya & Singh, Amrita & Lamba, Aditya & Gangwani, Naman & Soni,
Umang. (2019). Demand Forecasting Using Artificial Neural Networks—A Case
Study of American Retail Corporation. 10.1007/978-981-13-1822-1_8.
5. Yasir, M., Ansari, Y., Latif, K., Maqsood, H., Habib, A., Moon, J., & Rho, S.
(2022). Machine learning–assisted efficient demand forecasting using endogenous and
exogenous indicators for the textile industry. International Journal of Logistics
Research and Applications, 1–20. https://doi.org/10.1080/13675567.2022.2100334

31
6. Medina, Héctor & Pena, Mario & Siguenza-Guzman, Lorena & Guaman,
Rodrigo. (2022). Demand Forecasting for Textile Products Using Machine Learning
Methods.10.1007/978-3-031-03884-6_23.
https://www.researchgate.net/publication/359752767_Demand_Forecasting_for_Texti
le_Products_Using_Machine_Learning_Methods
7. Vashishtha, Rajesh & Burman, Vibhati & Kumar, Rajan & Sethuraman,
Srividhya & Sekar, Abhinaya & Ramanan, Sharadha. (2020). Product age based
demand forecast model for fashion retail.
https://www.researchgate.net/publication/342887066_Product_age_based_demand_fo
recast_model_for_fashion_retail
8. Lorente-Leyva, L.L., Alemany, M.M.E., Peluffo-Ordóñez, D.H., Araujo, R.A.
(2021). Demand Forecasting for Textile Products Using Statistical Analysis and
Machine Learning Algorithms. In: Nguyen, N.T., Chittayasothorn, S., Niyato, D.,
Trawiński, B. (eds) Intelligent Information and Database Systems. ACIIDS 2021.
Lecture Notes in Computer Science(), vol 12672. Springer, Cham.
https://doi.org/10.1007/978-3-030-73280-6_15

32