# A Method Based on GA-CNN-LSTM for Daily Tourist Flow Prediction at Scenic Spots

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

_{2.5}) CNN-LSTM. All those studies prove that CNN-LSTM model has excellent performance in prediction. On the basis of those studies, the current research scientifically selects the number of neurons of CNN-LSTM model by GA, which is scientific and efficient and makes CNN-LSTM model adaptive to predict tourist flow.

## 2. Related Methods

#### 2.1. Genetic Algorithm (GA)

#### 2.2. Convolutional Neural Network (CNN)

#### 2.3. Long-Short-Term Memory Network (LSTM)

_{t}), state memory cell (S

_{t−1}), and mid-output (h

_{t−1}) jointly determine the forgotten part of state memory cell in forgotten gate. x

_{t}determines the reserve vector in the state memory cell after sigmoid and tanh functions in input gate. Mid-output (h

_{t−1}) is determined by the updated (S

_{t}) and output (O

_{t}).

## 3. Method Based on GA-CNN-LSTM

_{conv}is the number of convolution kernels; n

_{1stm}is the number of neurons of LSTM; n

_{all}is the total number of convolution kernels of CNN and neurons of LSTM.

- Preliminarily selecting factors related to the scenic spot, such as historical tourist flow data, meteorological data, tickets data, and so on. Correlation analysis is performed on historical tourist flow data, meteorological data, and tickets data. Moreover, high correlation is selected as input.
- Selecting keywords of Baidu search index, which depends on what scenic spot the tourist is considering. Before traveling, tourists may search for information related to the scenic spot, such as weather, price, hotel, and so on.
- Performing a correlation analysis between the keywords of Baidu search index which are obtained in Step 2 and tourist flow; selecting keywords with higher correlation as input.
- Considering the lag of network search; setting a lag period and analyzing the correlation between Baidu index and tourist flow; choosing the lag period with the highest correlation.
- Constructing a new data set that is input into the GA-CNN-LSTM model for daily tourist flow prediction.
- Assessing accuracy of the GA-CNN-LSTM model; selecting relevant evaluation criteria for evaluation and comparing with related algorithms.

## 4. Empirical Study

#### 4.1. Data Set Construction

#### 4.2. Data Preprocessing

_{i}is the influencing factor of a certain tourist flow on day i, x

_{max}, x

_{min}are the maximum and minimum values of the data of the corresponding sequence.

#### 4.3. Data Set Partition

#### 4.4. Experimental Environment

#### 4.5. Model Building

_{i}is input value; y

_{i}is the output value after BN; m is the size of the mini-batch, that is, a mini-batch with m inputs; ${\mathsf{\mu}}_{\mathrm{B}}$ is the average of all inputs in the same mini-batch; ${\mathsf{\sigma}}_{\mathrm{B}}^{2}$ is the variance of all inputs in the same mini-batch; next, obtaining the normalized ${\widehat{\mathrm{x}}}_{\mathrm{i}}$ according to ${\mathsf{\mu}}_{\mathrm{B}}$, ${\mathsf{\sigma}}_{\mathrm{B}}^{2}$, ${\widehat{\mathrm{x}}}_{\mathrm{i}}$, and formula (12), putting ${\widehat{\mathrm{x}}}_{\mathrm{i}}$ into formula (13), and obtaining y

_{i}; $\gamma $ and $\mathsf{\beta}$ are obtained through machine learning. Using BN can maximize the neurons in deep neural networks to improve training efficiency.

## 5. Result and Discussion

_{avg}= 20.77. For 4 different CNN-LSTMs, MAPE

_{avg}= 23.08, 22.97, 22.89, 22.62, respectively. From the experimental results, the performance of GA-CNN-LSTM on MAPE can be concluded to be better than CNN-LSTM.

## 6. Conclusions

## Author Contributions

## Funding

## Acknowledgments

## Conflicts of Interest

## References

- Ministry of Culture and Tourism of the People’s Republic of China. Available online: http://zwgk.mct.gov.cn/ (accessed on 30 May 2019).
- Shannon, C.E. A mathematical theory of communication. Bell. Syst. Tech. J.
**1948**, 27, 379–423. [Google Scholar] [CrossRef] [Green Version] - Ponce-Flores, M.; Frausto-Solís, J.; Santamaría-Bonfil, G.; Pérez-Ortega, J.; González-Barbosa, J.J. Time Series Complexities and Their Relationship to Forecasting Performance. Entropy
**2020**, 22, 89. [Google Scholar] [CrossRef] [Green Version] - Guan, H.; Dai, Z.; Guan, S.; Zhao, A. A neutrosophic forecasting model for time series based on first-order state and information entropy of high-order fluctuation. Entropy
**2019**, 21, 455. [Google Scholar] [CrossRef] [Green Version] - Bretó, C.; Espinosa, P.; Hernández, P.; Pavía, J.M. An Entropy-Based Machine Learning Algorithm for Combining Macroeconomic Forecasts. Entropy
**2019**, 21, 1015. [Google Scholar] [CrossRef] [Green Version] - Sharda, N. Tourism Informatics: Visual Travel Recommender Systems, Social Communities, and User Interface Design; Information Science Reference: New York, NY, USA, 2010. [Google Scholar]
- Zi, L.U.; Liu, N.; Zui, Z. The guiding effect of information flow of Australian tourism website on tourist flow: Process, intensity and mechanism. Hum. Geogr.
**2007**, 22, 88–93. [Google Scholar] - Gustavsson, P.; Nordström, J. The impact of seasonal unit roots and vector ARMA modelling on forecasting monthly tourism flows. Tour. Econ.
**2001**, 7, 117–133. [Google Scholar] [CrossRef] - Lim, C.; McAleer, M. Monthly seasonal variations: Asian tourism to Australia. Ann. Tour. Res.
**2001**, 28, 68–82. [Google Scholar] [CrossRef] - Lim, C.; McAleer, M. A seasonal analysis of Asian tourist arrivals to Australia. Appl. Econ.
**2000**, 32, 499–509. [Google Scholar] [CrossRef] - Kim, S.; Song, H. Analysis of inbound tourism demand in South Korea: A cointegration and error correction approach. Tour. Anal.
**1998**, 3, 25–41. [Google Scholar] - Ao, S. A hybrid neural network cybernetic system for quantifying cross-market dynamics and business forecasting. Soft Comput.
**2011**, 15, 1041–1053. [Google Scholar] [CrossRef] - Lin, V.S.; Liu, A.; Song, H. Modeling and forecasting Chinese outbound tourism: An econometric approach. J. Travel Tour. Mark.
**2015**, 32, 34–49. [Google Scholar] [CrossRef] - Chen, C.-F.; Lai, M.-C.; Yeh, C.-C. Forecasting tourism demand based on empirical mode decomposition and neural network. Knowl. Based Syst.
**2012**, 26, 281–287. [Google Scholar] [CrossRef] - Law, R. Back-propagation learning in improving the accuracy of neural network-based tourism demand forecasting. Tour. Manag.
**2000**, 21, 331–340. [Google Scholar] [CrossRef] - Pai, P.-F.; Hung, K.-C.; Lin, K.-P. Tourism demand forecasting using novel hybrid system. Expert Syst. Appl.
**2014**, 41, 3691–3702. [Google Scholar] [CrossRef] - Sun, S.; Wei, Y.; Tsui, K.-L.; Wang, S. Forecasting tourist arrivals with machine learning and internet search index. Tour. Manag.
**2019**, 70, 1–10. [Google Scholar] [CrossRef] - Li, X.; Pan, B.; Law, R.; Huang, X. Forecasting tourism demand with composite search index. Tour. Manag.
**2017**, 59, 57–66. [Google Scholar] [CrossRef] - Chen, R.; Liang, C.-Y.; Hong, W.-C.; Gu, D.-X. Forecasting holiday daily tourist flow based on seasonal support vector regression with adaptive genetic algorithm. Appl. Soft Comput.
**2015**, 26, 435–443. [Google Scholar] [CrossRef] - Li, K.; Lu, W.; Liang, C.; Wang, B. Intelligence in Tourism Management: A Hybrid FOA-BP Method on Daily Tourism Demand Forecasting with Web Search Data. Mathematics
**2019**, 7, 531. [Google Scholar] [CrossRef] [Green Version] - Song, G.F.; Liang, C.Y.; Liang, Y.; Zhao, S.P.; Management, S.O. Prediction for Tourist Attractions Daily Traffic Based on Modified Genetic Algorithm Optimized BP Neural Network. J. Chin. Comput. Syst.
**2014**, 20, 232–238. [Google Scholar] - Li, Y.; Cao, H. Prediction for tourism flow based on LSTM neural network. Procedia Comput. Sci.
**2018**, 129, 277–283. [Google Scholar] [CrossRef] - Khalid, R.; Javaid, N.; Al-zahrani, F.A.; Aurangzeb, K.; Qazi, E.-U.-H.; Ashfaq, T. Electricity Load and Price Forecasting Using Jaya-Long Short Term Memory (JLSTM) in Smart Grids. Entropy
**2020**, 22, 10. [Google Scholar] [CrossRef] [Green Version] - Wei, W.; Jia, X.; Liu, Y.; Yu, X. Travel Time Forecasting with Combination of Spatial-Temporal and Time Shifting Correlation in CNN-LSTM Neural Network. In Asia-Pacific Web (APWeb) and Web-Age Information Management (WAIM) Joint International Conference on Web and Big Data; Springer: Cham, Switzerland, 2018; pp. 297–311. [Google Scholar]
- Shen, Y.; Hua, J.; Jin, C.; Huang, D. TCL: Tensor-CNN-LSTM for Travel Time Prediction with Sparse Trajectory Data. In International Conference on Database Systems for Advanced Applications; Springer: Cham, Switzerland, 2019; pp. 329–333. [Google Scholar]
- Kim, T.-Y.; Cho, S.-B. Predicting residential energy consumption using CNN-LSTM neural networks. Energy
**2019**, 182, 72–81. [Google Scholar] [CrossRef] - Chen, R.; Wang, X.; Zhang, W.; Zhu, X.; Li, A.; Yang, C. A hybrid CNN-LSTM model for typhoon formation forecasting. Geoinformatica
**2019**, 23, 375–396. [Google Scholar] [CrossRef] - Huang, C.-J.; Kuo, P.-H. A deep cnn-lstm model for particulate matter (PM 2.5) forecasting in smart cities. Sensors
**2018**, 18, 2220. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Holland, J.H. Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence; MIT Press: Cambridge, MA, USA, 1992. [Google Scholar]
- Dejong, K. An Analysis of the Behavior of a Class of Genetic Adaptive Systems. Ph.D. Thesis, University of Michigan, Ann Arbor, MI, USA, 1975. [Google Scholar]
- LeCun, Y.; Bottou, L.; Bengio, Y.; Haffner, P. Gradient-based learning applied to document recognition. Proc. IEEE
**1998**, 86, 2278–2324. [Google Scholar] [CrossRef] [Green Version] - Hochreiter, S.; Schmidhuber, J. Long short-term memory. Neural Comput.
**1997**, 9, 1735–1780. [Google Scholar] [CrossRef] - Nadarajah, S.; Afuecheta, E.; Chan, S. GARCH modeling of five popular commodities. Empir. Econ.
**2015**, 48, 1691–1712. [Google Scholar] [CrossRef] - Deng, N.-Y.; Tian, Y. A New Method of Data Mining-Support Vector Machine; Science Publication: Beijing, China, 2004; Volume 51. [Google Scholar]
- Pelckmans, K.; Suykens, J.A.; De Moor, B. Regularization constants in LS-SVMs: A fast estimate via convex optimization. In Proceedings of the 2004 IEEE International Joint Conference on Neural Networks (IEEE Cat. No. 04CH37541), Budapest, Hungary, 25–29 July 2004; pp. 699–704. [Google Scholar]
- Klambauer, G.; Unterthiner, T.; Mayr, A.; Hochreiter, S. Self-normalizing neural networks. In Proceedings of the Advances in Neural Information Processing Systems 30, Long Beach, CL, USA, 4–9 December 2017. [Google Scholar]
- Ioffe, S.; Szegedy, C. Batch normalization: Accelerating deep network training by reducing internal covariate shift. Arxiv
**2015**, arXiv:1502.03167v3. [Google Scholar]

Corresponding Rules | Yesterday | The Day before Yesterday | Same Day Last Week | 30 Days Ago | 365 Days Ago |
---|---|---|---|---|---|

Correlation | 0.710 | 0.433 | 0.428 | 0.026 | 0.410 |

Lag Period | 1 | 2 | 3 | 4 | 5 | 5 | 7 |
---|---|---|---|---|---|---|---|

Correlation | 0.716 | 0.504 | 0.364 | 0.311 | 0.256 | 0.255 | 0.320 |

Keywords | Correlation |
---|---|

Huangshan Scenic Spot | 0.552 |

Huangshan Travel Guide | 0.523 |

Huangshan Hong Village | 0.636 |

Huangshan Travel Map | 0.499 |

Huangshan Day Tour | 0.585 |

Lag Period | Correlation with the Different Lag Period | ||||
---|---|---|---|---|---|

Huangshan Scenic Spot | Huangshan Travel Guide | Huangshan Hong Village | Huangshan Travel Map | Huangshan Day Tour | |

1 | 0.578 | 0.564 | 0.627 | 0.564 | 0.651 |

2 | 0.607 | 0.579 | 0.640 | 0.591 | 0.653 |

3 | 0.593 | 0.573 | 0.603 | 0.567 | 0.596 |

4 | 0.575 | 0.556 | 0.569 | 0.534 | 0.551 |

5 | 0.569 | 0.543 | 0.556 | 0.516 | 0.522 |

6 | 0.519 | 0.516 | 0.504 | 0.458 | 0.496 |

7 | 0.461 | 0.303 | 0.465 | 0.409 | 0.465 |

15 | 0.427 | 0.436 | 0.415 | 0.374 | 0.398 |

31 | 0.383 | 0.351 | 0.296 | 0.206 | 0.292 |

Impact Factors | Characteristics |
---|---|

Tourist flow related historical data: | The number of tourists yesterday; The number of tourists the day before yesterday; The number of tourists 365 days ago; The number of tourists Same day last week; The number of tickets. |

Time factors: | Monday to Sunday; Holiday or Working day. |

Meteorological factors: | Weather; Wind speed; Average temperature; Average humidity. |

Baidu search index: | Huangshan Scenic Spot |

Huangshan Travel Guide | |

Huangshan Hong Village | |

Huangshan Travel Map | |

Huangshan Day Tour |

CNN-LSTM (1) | CNN-LSTM (2) | CNN-LSTM (3) | CNN-LSTM (4) |
---|---|---|---|

64 | 32 | 64 | 32 |

128 | 64 | 128 | 64 |

256 | 128 | 256 | 128 |

16 | 32 | 32 | 16 |

32 | 64 | 64 | 32 |

64 | 128 | 128 | 64 |

Test | MAPE (%) | ||||
---|---|---|---|---|---|

GA-CNN -LSTM | CNN-LSTM (1) | CNN-LSTM (2) | CNN-LSTM (3) | CNN-LSTM (4) | |

1 | 20.73 | 22.96 | 23.71 | 22.52 | 22.90 |

2 | 20.50 | 22.93 | 22.71 | 22.77 | 22.29 |

3 | 20.86 | 23.71 | 22.89 | 23.80 | 22.56 |

4 | 20.79 | 22.62 | 23.10 | 22.96 | 22.64 |

5 | 20.96 | 23.18 | 22.43 | 22.41 | 22.74 |

Average | 20.77 | 23.08 | 22.97 | 22.89 | 22.62 |

Test | GA-CNN -LSTM | CNN-LSTM | LSTM | CNN | BP |
---|---|---|---|---|---|

1 | 20.73 | 22.90 | 24.92 | 29.81 | 28.36 |

2 | 20.50 | 22.29 | 23.96 | 29.81 | 28.56 |

3 | 20.86 | 22.56 | 26.64 | 29.80 | 28.33 |

4 | 20.79 | 22.64 | 24.54 | 29.81 | 29.02 |

5 | 20.96 | 22.74 | 24.56 | 29.81 | 27.18 |

Average | 20.77 | 22.63 | 24.92 | 29.81 | 28.29 |

Test | GA-CNN -LSTM | CNN-LSTM | LSTM | CNN | BP |
---|---|---|---|---|---|

1 | 0.911 | 0.908 | 0.847 | 0.887 | 0.875 |

2 | 0.912 | 0.900 | 0.842 | 0.887 | 0.887 |

3 | 0.912 | 0.908 | 0.847 | 0.887 | 0.889 |

4 | 0.916 | 0.901 | 0.846 | 0.887 | 0.874 |

5 | 0.912 | 0.905 | 0.837 | 0.885 | 0.903 |

Average | 0.913 | 0.904 | 0.844 | 0.887 | 0.886 |

Test | GA-CNN -LSTM | CNN-LSTM | LSTM | CNN | BP |
---|---|---|---|---|---|

1 | 0.923 | 0.929 | 0.915 | 0.901 | 0.893 |

2 | 0.922 | 0.919 | 0.911 | 0.901 | 0.889 |

3 | 0.919 | 0.912 | 0.920 | 0.906 | 0.902 |

4 | 0.921 | 0.910 | 0.917 | 0.906 | 0.902 |

5 | 0.904 | 0.913 | 0.918 | 0.906 | 0.861 |

Average | 0.919 | 0.917 | 0.916 | 0.904 | 0.889 |

Month | GA-CNN-LSTM | CNN-LSTM | LSTM | CNN | BP |
---|---|---|---|---|---|

1 | 24.9741 | 36.0873 | 46.7268 | 52.3842 | 62.7581 |

2 | 33.9763 | 36.3812 | 26.212 | 33.9085 | 33.4193 |

3 | 32.0348 | 33.182 | 32.3281 | 42.7591 | 23.8872 |

4 | 27.414 | 25.944 | 22.537 | 27.274 | 31.5567 |

5 | 15.8333 | 17.3235 | 16.4253 | 28.1739 | 22.1647 |

6 | 14.5825 | 15.5871 | 16.516 | 25.1302 | 22.4283 |

7 | 9.5686 | 8.4264 | 12.4824 | 18.8382 | 19.6289 |

8 | 15.7049 | 13.2258 | 14.6972 | 19.6488 | 12.7721 |

9 | 19.3528 | 16.9775 | 16.4402 | 20.7455 | 25.8335 |

10 | 15.0045 | 20.249 | 26.1285 | 29.4281 | 31.2337 |

11 | 14.8754 | 19.8404 | 18.0203 | 29.9235 | 24.6147 |

12 | 23.8101 | 25.2183 | 38.4586 | 29.3431 | 32.626 |

Season | GA-CNN-LSTM | CNN-LSTM | LSTM | CNN | BP |
---|---|---|---|---|---|

1 | 30.3284 | 35.2168 | 35.089 | 43.0172 | 40.0215 |

2 | 19.2766 | 19.6182 | 18.4928 | 26.8593 | 25.3832 |

3 | 14.8754 | 12.8766 | 14.5399 | 19.7442 | 19.4115 |

4 | 17.8741 | 21.7692 | 27.5358 | 29.5649 | 29.4915 |

© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Lu, W.; Rui, H.; Liang, C.; Jiang, L.; Zhao, S.; Li, K.
A Method Based on GA-CNN-LSTM for Daily Tourist Flow Prediction at Scenic Spots. *Entropy* **2020**, *22*, 261.
https://doi.org/10.3390/e22030261

**AMA Style**

Lu W, Rui H, Liang C, Jiang L, Zhao S, Li K.
A Method Based on GA-CNN-LSTM for Daily Tourist Flow Prediction at Scenic Spots. *Entropy*. 2020; 22(3):261.
https://doi.org/10.3390/e22030261

**Chicago/Turabian Style**

Lu, Wenxing, Haidong Rui, Changyong Liang, Li Jiang, Shuping Zhao, and Keqing Li.
2020. "A Method Based on GA-CNN-LSTM for Daily Tourist Flow Prediction at Scenic Spots" *Entropy* 22, no. 3: 261.
https://doi.org/10.3390/e22030261