Demand Forecasting for Multichannel Fashion Retailers by Integrating Clustering and Machine Learning Algorithms
Abstract
:1. Introduction
2. Literature Review
2.1. Demand Forecasting in the Fashion Industry
2.2. K-Means Clustering
2.3. Extreme Learning Machines
2.4. Support Vector Regression
3. Proposed Clustering-Based Demand Forecasting Model
3.1. Proposed Scheme
- (1)
- Data collection: Collect raw data and divide those data into training and testing data sets with a ratio of approximately 9:1.
- (2)
- Data clustering: Group the training data through the KM algorithm.
- (3)
- Cluster assignment: Calculate the Euclidean distance between each point of the test data and each training data cluster center and determine the training data cluster corresponding to each test datum by finding the shortest distance to make predictions.
- (4)
- Model building, assessment, and assignment: Combine the machine learning method ELM or SVR to establish a prediction model for each cluster. Determine the best forecasting technology for each group according to performance indicators.
- (5)
- Explainability: Propose an explanation based on the results of the KM-ELM and KM-SVR prediction models.
3.2. Collecting and Processing Data
3.3. Performance Evaluation Metrics
- (1)
- MAPE:
- (2)
- RMSE:
4. Empirical Analysis
4.1. Empirical Data
4.2. Predictor Variables
4.3. Results
4.3.1. Clustering-Based Prediction Model
4.3.2. Comparison of Prediction Models
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- Frings, G.S. Fashion: From Concept to Consumer; Prentice-Hall: Englewood Cliffs, NJ, USA, 1982; p. 305. [Google Scholar]
- Nenni, M.E.; Giustiniano, L.; Pirolo, L. Demand Forecasting in the Fashion Industry: A Review. Int. J. Eng. Bus. Manag. 2013, 5, 37. [Google Scholar] [CrossRef] [Green Version]
- Yu, Y.; Choi, T.-M.; Hui, C.-L. An intelligent fast sales forecasting model for fashion products. Expert Syst. Appl. 2010, 38, 7373–7379. [Google Scholar] [CrossRef]
- Hossain, M.M.; Abdulla, F. Forecasting the garlic production in Bangladesh by ARIMA Model. Asian J. Crop Sci. 2015, 7, 147. [Google Scholar]
- Hsu, C.-C.; Chen, C.-Y. Applications of improved grey prediction model for power demand forecasting. Energy Convers. Manag. 2003, 44, 2241–2249. [Google Scholar] [CrossRef]
- Choi, T.-M.; Hui, C.-L.; Liu, N.; Ng, S.-F.; Yu, Y. Fast fashion sales forecasting with limited data and time. Decis. Support Syst. 2014, 59, 84–92. [Google Scholar] [CrossRef]
- Sun, Z.L.; Choi, T.M.; Au, K.F.; Yu, Y. Sales forecasting using extreme learning machine with applications in fashion re-tailing. Decis. Support Syst. 2008, 46, 411–419. [Google Scholar] [CrossRef]
- Lu, C.J.; Shao, Y.E. Forecasting computer products sales by integrating ensemble empirical mode decomposition and extreme learning machine. Math. Probl. Eng. 2012, 2012, 831201. [Google Scholar] [CrossRef]
- Lu, C.-J. Sales forecasting of computer products based on variable selection scheme and support vector regression. Neurocomputing 2014, 128, 491–499. [Google Scholar] [CrossRef]
- Yu, L.; Dai, W.; Tang, L. A novel decomposition ensemble model with extended extreme learning machine for crude oil price forecasting. Eng. Appl. Artif. Intell. 2016, 47, 110–121. [Google Scholar] [CrossRef]
- Chen, I.-F.; Lu, C.-J. Sales forecasting by combining clustering and machine-learning techniques for computer retailing. Neural Comput. Appl. 2016, 28, 2633–2647. [Google Scholar] [CrossRef]
- Tan, J.Y.B.; Bong, D.B.L.; Rigit, A.R.H. Time series prediction using backpropagation network optimized by Hybrid K-means-Greedy Algorithm. Eng. Lett. 2012, 20, 203–210. [Google Scholar]
- Benmouiza, K.; Cheknane, A. Forecasting hourly global solar radiation using hybrid K-means and nonlinear autoregressive neural network models. Energy Convers. Manag. 2013, 75, 561–569. [Google Scholar] [CrossRef]
- Lu, C.J.; Lee, T.S.; Lian, C.M. Sales forecasting for computer wholesalers: A comparison of multivariate adaptive regression splines and artificial neural networks. Decis. Support Syst. 2012, 54, 584–596. [Google Scholar] [CrossRef]
- Ramos, P.; Santos, N.; Rebelo, R. Performance of state space and ARIMA models for consumer retail sales forecasting. Robot. Comput. Manuf. 2015, 34, 151–163. [Google Scholar] [CrossRef] [Green Version]
- Murat, M.; Malinowska, I.; Gos, M.; Krzyszczak, J. Forecasting daily meteorological time series using ARIMA and regression models. Int. Agrophysics 2018, 32, 253–264. [Google Scholar] [CrossRef]
- Chen, S.; Chen, J. Forecasting container throughput at ports using genetic programming. Expert Syst. Appl. 2010, 37, 2054–2058. [Google Scholar] [CrossRef]
- Frank, C.; Garg, A.; Sztandera, L.; Raheja, A. Forecasting women’s apparel sales using mathematical modeling. Int. J. Cloth. Sci. Technol. 2003, 15, 107–125. [Google Scholar] [CrossRef] [Green Version]
- Au, K.-F.; Choi, T.-M.; Yu, Y. Fashion retail forecasting by evolutionary neural networks. Int. J. Prod. Econ. 2008, 114, 615–630. [Google Scholar] [CrossRef]
- MacQueen, J. Some methods for classification and analysis of multivariate observations. In Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, Berkeley, CA, USA, 21 June–18 July 1965; Volume 1, pp. 281–297. [Google Scholar]
- Guftar, M.; Ali, S.H.; Raja, A.A.; Qamar, U. A novel framework for classification of syncope disease using K-means clustering algorithm. In Proceedings of the 2015 SAI Intelligent Systems Conference (IntelliSys), London, UK, 10–11 November 2015; pp. 127–132. [Google Scholar]
- Xue, W.; Feijia, L.; Wenxia, X.; Kun, G.; Guodong, L. Based on K-Means Clustering and CNN Algorithm Research in Hail Cloud Determination. In Proceedings of the 2015 Seventh International Conference on Measuring Technology and Mechatronics Automation, Nanchang, China, 13–14 June 2015; pp. 232–235. [Google Scholar]
- Huang, G.-B.; Chen, L.; Siew, C.-K. Universal Approximation Using Incremental Constructive Feedforward Networks With Random Hidden Nodes. IEEE Trans. Neural Netw. 2006, 17, 879–892. [Google Scholar] [CrossRef] [Green Version]
- Benoît, F.; van Heeswijk, M.; Miche, Y.; Verleysen, M.; Lendasse, A. Feature selection for nonlinear models with extreme learning machines. Neurocomputing 2013, 102, 111–124. [Google Scholar] [CrossRef]
- Huang, G.; Song, S.; Gupta, J.N.D.; Wu, C. Semi-Supervised and Unsupervised Extreme Learning Machines. IEEE Trans. Cybern. 2014, 44, 2405–2417. [Google Scholar] [CrossRef]
- Kasun, L.L.C.; Zhou, H.; Huang, G.B.; Vong, C.M. Representational learning with extreme learning machine for big data. IEEE Intell. Syst. 2013, 28, 31–34. [Google Scholar]
- Huang, G.; Huang, G.B.; Song, S.; You, K. Trends in extreme learning machines: A review. Neural Netw. 2015, 61, 32–48. [Google Scholar] [CrossRef] [PubMed]
- Xia, M.; Lu, W.; Yang, J.; Ma, Y.; Yao, W.; Zheng, Z. A hybrid method based on extreme learning machine and k-nearest neighbor for cloud classification of ground-based visible cloud image. Neurocomputing 2015, 160, 238–249. [Google Scholar] [CrossRef]
- Huang, G.-B.; Zhu, Q.-Y.; Siew, C.-K. Extreme learning machine: Theory and applications. Neurocomputing 2006, 70, 489–501. [Google Scholar] [CrossRef]
- Shrivastava, N.A.; Panigrahi, B.K. A hybrid wavelet-ELM based short term price forecasting for electricity markets. Int. J. Electr. Power Energy Syst. 2014, 55, 41–50. [Google Scholar] [CrossRef]
- Vapnik, V.N. The Nature of Statistical Learning Theory; Springer, Inc.: New York, NY, USA, 1995. [Google Scholar]
- Yu, X.; Qi, Z.; Zhao, Y. Support Vector Regression for Newspaper/Magazine Sales Forecasting. Procedia Comput. Sci. 2013, 17, 1055–1062. [Google Scholar] [CrossRef] [Green Version]
- Du, X.F.; Leung, S.C.; Zhang, J.L.; Lai, K.K. Demand forecasting of perishable farm products using support vector machine. Int. J. Syst. Sci. 2013, 44, 556–567. [Google Scholar] [CrossRef]
- Gunn, S.R. Support vector machines for classification and regression. ISIS Tech. Rep. 1998, 14, 5–16. [Google Scholar]
- Hsu, C.W.; Chang, C.C.; Lin, C.J. A Practical Guide to Support Vector Classification; National Taiwan University: Taipei, Taiwan, 2003; pp. 1–16. [Google Scholar]
- Cai, X.; Nan, X.Y.; Gao, B.P. Oxygen supply prediction model based on IWO-SVR in bio-oxidation pretreatment. Eng. Lett. 2015, 23, 173–179. [Google Scholar]
- Kim, S.; Kim, H. A new metric of absolute percentage error for intermittent demand forecasts. Int. J. Forecast. 2016, 32, 669–679. [Google Scholar] [CrossRef]
- Willmott, C.; Matsuura, K. Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance. Clim. Res. 2005, 30, 79–82. [Google Scholar] [CrossRef]
- Lewis, E.B. Control of body segment differentiation in drosophila by the bithorax gene complex. Prog. Clin. Biol. Res. 1982, 1, 269–288. [Google Scholar]
- Lu, C.J.; Chen, I.F. Identifying important predictors for computer server sales using an effective hybrid forecasting technique. Int. J. Inf. Manag. Sci. 2017, 28, 213–232. [Google Scholar]
- Lu, C.J.; Kao, L.J. A clustering-based sales forecasting scheme by using extreme learning machine and ensembling link-age methods with applications to computer server. Eng. Appl. Artif. Intell. 2016, 55, 231–238. [Google Scholar] [CrossRef]
- Jain, A.K. Data clustering: 50 years beyond K-means. Pattern Recognit. Lett. 2010, 31, 651–666. [Google Scholar] [CrossRef]
Number of Physical Stores | Prefecture | Region | City |
---|---|---|---|
108 | Tokyo | Kantō | Tokyo |
73 | Osaka | Kansai | Osaka |
48 | Aichi | Chūbu | Nagoya |
33 | Fukuoka | Kyūshū·Okinawa | Fukuoka |
29 | Hokkaido | Hokkaidō·Tōhoku | Sapporo |
19 | Hiroshima | Shikoku·Chūgoku | Hiroshima |
Issue | Issue 1 | Issue 2 | Issue 3 | Issue 4 | |
---|---|---|---|---|---|
Variables | Physical Stores Demand | Physical Stores Demand with Meteorological Data | Non-store Demand | Non-Store Demand with Meteorological Data | |
Endogenous variables | (X1) Rate of change in the last month (t-1) | X1_P | X1_P | X1_N | X1_N |
(X2) Rate of change in the last two months (t-2) | X2_P | X2_P | X2_N | X2_N | |
(X3) Rate of change in the last six months (t-6) | X3_ P | X3_ P | X3_N | X3_N | |
(X4) Rate of change of the same month of last year (t-12) | X4_ P | X4_ P | X4_N | X4_N | |
(X5) Moving average of the rate of change in the past two months (MA2) | X5_p | X5_ P | X5_N | X5_N | |
(X6) Moving average of the rate of change in the past three months (MA3) | X6_P | X6_P | X6_N | X6_N | |
(X7) Moving average of the rate of change in the past six months (MA6) | X7_P | X7_P | X7_N | X7_N | |
(X8) BIAS of the last two months (BIAS2) | X8_P | X8_P | X8_N | X8_N | |
(X9) BIAS of the rate of change of the last three months (BIAS3) | X9_P | X9_P | X9_N | X9_N | |
(X10) BIAS of the rate of change of the same store of the last six months (BIAS6) | X10_P | X10_P | X10_N | X10_N | |
Exogenous variable | (X11) Average temperature of Sapporo | X11 | X11 | ||
(X12) Average temperature of Nagoya | X12 | X12 | |||
(X13) Average temperature of Tokyo | X13 | X13 | |||
(X14) Average temperature of Osaka | X14 | X14 | |||
(X15) Average temperature of Hiroshima | X15 | X15 | |||
(X16) Average temperature of Fukuoka | X16 | X16 | |||
(X17) Average temperate of all six regions (average of X11 to X16) | X17 | X17 | |||
(X18) Average rainfall of Sapporo | X18 | X18 | |||
(X19) Average rainfall of Nagoya | X19 | X19 | |||
(X20) Average rainfall of Tokyo | X20 | X20 | |||
(X21) Average rainfall of Osaka | X21 | X21 | |||
(X22) Average rainfall of Hiroshima | X22 | X22 | |||
(X23) Average rainfall of Fukuoka | X23 | X23 | |||
(X24) Average rainfall of all six regions (average of X18 to X23) | X24 | X24 |
Activation Function | Number of Hidden Nodes | MAPE | RMSE |
---|---|---|---|
Linear Transfer Function | 10 | 0.44% | 0.62 |
18 | 0.44% | 0.62 | |
19 | 0.44% | 0.62 | |
20 | 0.44% | 0.62 | |
21 | 0.44% | 0.62 | |
22 | 0.44% | 0.62 | |
Sigmoid Transfer Function | 10 | 9.11% | 11.21 |
18 | 7.60% | 9.96 | |
19 | 7.96% | 10.53 | |
20 | 8.27% | 11.01 | |
21 | 9.02% | 11.32 | |
22 | 7.29% | 9.52 | |
Radial Basis Transfer Function | 10 | 57.20% | 73.39 |
18 | 99.99% | 106.75 | |
19 | 91.97% | 102.20 | |
20 | 85.82% | 96.05 | |
21 | 77.47% | 89.30 | |
22 | 88.61% | 100.04 |
MAPE | RMSE | |||||
---|---|---|---|---|---|---|
Activation Function/Number of Clusters | Linear | Sigmoid | Radial Basis | Linear | Sigmoid | Radial Basis |
2 | 0.31% | 6.72% | 58.10% | 0.38 | 9.02 | 73.80 |
3 | 0.25% | 6.57% | 60.55% | 0.30 | 9.19 | 74.90 |
4 | 0.22% | 5.73% | 34.42% | 0.27 | 7.5 | 46.80 |
5 | 0.13% | 4.79% | 42.68% | 0.14 | 5.82 | 50.58 |
Cluster 1 | Cluster 2 | Cluster3 | Cluster 4 | Cluster 5 | Total | |
---|---|---|---|---|---|---|
Training data | 24 | 66 | 35 | 11 | 8 | 144 |
Testing data | 2 | 5 | 3 | 1 | 1 | 12 |
Total | 26 | 71 | 38 | 12 | 9 | 156 |
MAPE | RMSE | |||
---|---|---|---|---|
Kernel Function/Number of Clusters | Radial | Linear | Radial | Linear |
2 | 4.13% | 0.33% | 5.3 | 0.42 |
3 | 5.47% | 0.21% | 7.38 | 0.29 |
4 | 5.20% | 0.24% | 6.13 | 0.30 |
5 | 5.69% | 0.23% | 6.70 | 0.26 |
Cluster 1 | Cluster 2 | Cluster 3 | Total | |
---|---|---|---|---|
Training data | 61 | 54 | 29 | 144 |
Testing data | 5 | 5 | 2 | 12 |
Total | 66 | 59 | 31 | 156 |
ELM | SVR | KM-ELM | KM-SVR | ||
---|---|---|---|---|---|
10-predictor forecasting model (Issue 1) | MAPE | 0.44% | 0.42% | 0.13%↓ | 0.21%↓ |
RMSE | 0.62 | 0.59 | 0.14↓ | 0.29↓ | |
24-predictor forecasting model (Issue 2) | MAPE | 0.70% | 0.68% | 0.50%↓ | 0.57%↓ |
RMSE | 0.80 | 0.73 | 0.64↓ | 0.68↓ |
ELM | SVR | KM-ELM | KM-SVR | ||
---|---|---|---|---|---|
10-predictor forecasting model (Issue 3) | MAPE | 0.32% | 0.42% | 0.14%↓ | 0.26%↓ |
RMSE | 0.41 | 0.40 | 0.15↓ | 0.21↓ | |
24-predictor forecasting model (Issue 4) | MAPE | 0.31% | 0.36% | 0.25%↓ | 0.27%↓ |
RMSE | 0.40 | 0.40 | 0.23↓ | 0.22↓ |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Chen, I.-F.; Lu, C.-J. Demand Forecasting for Multichannel Fashion Retailers by Integrating Clustering and Machine Learning Algorithms. Processes 2021, 9, 1578. https://0-doi-org.brum.beds.ac.uk/10.3390/pr9091578
Chen I-F, Lu C-J. Demand Forecasting for Multichannel Fashion Retailers by Integrating Clustering and Machine Learning Algorithms. Processes. 2021; 9(9):1578. https://0-doi-org.brum.beds.ac.uk/10.3390/pr9091578
Chicago/Turabian StyleChen, I-Fei, and Chi-Jie Lu. 2021. "Demand Forecasting for Multichannel Fashion Retailers by Integrating Clustering and Machine Learning Algorithms" Processes 9, no. 9: 1578. https://0-doi-org.brum.beds.ac.uk/10.3390/pr9091578