A Prediction Model for Hallabong Tangor Product Prices using LSTM (Long Short-term Memory) Network

Dae Ho Jung; Young-Yeol Cho

doi:10.7235/HORT.20220051

All Issue

2022 Vol.40, Issue 5 Preview Page

Research Article

A Prediction Model for Hallabong Tangor Product Prices using LSTM (Long Short-term Memory) Network LSTM 네트워크를 이용한 한라봉 가격 예측 모델

31 October 2022. pp. 571-577

PDF XML

Abstract

The predicted prices of agricultural products have a great influence on market supply and demand control decisions. The purpose of this study was to predict the average price of Hallabong tangor, which is distributed in 9 wholesale markets nationwide, using a deep learning-based LSTM network. Daily data (2,398 data points) from 2007 to January 2021 based on the average price of 3 kg of Hallabong tangor sold in 9 wholesale markets were used in this study. To train the model, 80% of the dataset was used as training data and 20% was used as test data. For accurate price predictions using an LSTM network, it is necessary to determine the epoch and batch size. An epoch of 700 and batch size of 128 were judged to be suitable for the model (RMSE = 1040). The resulting model could accurately predict the average price of Hallabong tangor. The deep learning-based Hallabong tangor price prediction model provides useful information for agricultural product price trends, government policy establishment, and consumer demands.

Keywords

artificial intelligence

deep learning

long-term dependency

recurrent neural network

농산물의 가격 예측은 시장 수급 조절 결정에 큰 영향을 끼친다. 본 연구의 목적은 전국 9개 도매시장에서 유통되고 있는 한라봉의 평균가격을 딥러닝 기반의 LSTM 네트워크를 이용하여 한라봉 가격을 예측하고자 하였다. 9개의 도매시장에서 거래되고 있는 한라봉의 3kg 단위 평균가격을 바탕으로 2007년부터 2021년 1월까지 2,398개의 일일 데이터를 사용하였다. Training 데이터는 전체 데이터의 80%, test 데이터는 전체 데이터의 20%를 사용하여 모델을 학습시켰다. LSTM 네트워크를 위해 보다 더 정확한 가격을 예측하기 위해서는 epoch와 batch size를 결정해야 하는데, epoch는 700과 batch size 128가 적합한 모델이라 판단하였다. 이 값을 LSTM 네트워크를 적용한 결과, 비교적 한라봉의 평균 가격을 정확하게 예측할 수 있었다(RMSE = 1040). 딥러닝 기반 한라봉 가격 예측 모델은 농산물 가격 동향과 정부 정책 수립 및 소비자들 요구에게도 유용한 정보로 활용될 수 있을 것이다.

키워드

딥러닝

인공지능

순환신경망

장기 의존성

References

Banerjee T, Sinha S, Choudhury P (2022) Long term and short term forecasting of horticultural produce based on the LSTM network model. Appl Intell 52:9117-9147. doi:10.1007/s10489-021-02845-x 10.1007/s10489-021-02845-x

Han SH, Kang JH (2011) Effect of water states of fruit vesicle and leaf on fruit quality in 'Trifoliate' orange and 'Swingle citrumelo' rootstock of 'Shiranuhi' [(Citrus unshiu × C. sinensis) × C. reticulata] mandarin Hybrid, M16A line in plastic film house cultivation. J Bio-Env Con 20:204-210

Hochreiter S, Schmidhuber J (1997) Long short-term memory. Neural Comput 9:1735-1780. doi:10.1162/neco.1997.9.8.1735 10.1162/neco.1997.9.8.17359377276

Im JM, Kim WY, Byoun WJ, Shin SJ (2018) Fruit price prediction study using artificial intelligence. JCCT 4:197-204. doi:10.17703/JCCT.2018.4.2.197 10.17703/JCCT.2018.4.2.197

Kim BS (2005) A comparison on forecasting performance of the application models for Forecasting of vegetable prices. KJAT 46:89-113

Kim SK, Oh TI (2018) Real-time PM10 concentration prediction LSTM model based on IoT streaming sensor data. JKAIS 19:310-318. doi:10.5762/KAIS.2018.19.11.310 10.5762/KAIS.2018.19.11.310

Ko SB (2017) A study on the analysis of shipment structure of hallabong tangor in Korea. JKAIS 18:670-676. doi:10.5762/KAIS.2017.18.7.670 10.5762/KAIS.2017.18.7.670

Ko SB, Kim BS (2012) A study on the model specification for supply-demand forecast of hallabong tangor in Korea. JKAIS 13:5163-5168. doi:10.5762/KAIS.2012.13.11.5163 10.5762/KAIS.2012.13.11.5163

Kurumatani K (2020) Time series forecasting of agricultural product prices based on recurrent neural networks and its evaluation method. SN Appl Sci 2:1434. doi:10.1007/s42452-020-03225-9 10.1007/s42452-020-03225-9

Lee JK, Oh JW, Cho YJ, Lee DH (2020) A research about time domain estimation method for greenhouse environmental factors based on artificial intelligence. Protected Hortic Plant Fac 29:277-284. doi:10.12791/KSBEC.2020.29.3.277 10.12791/KSBEC.2020.29.3.277

Lee YS, Choi BO, Shim SB (2006) A time series analysis on prices of fruits and vegetables. Agri-Policy Focus R537, KREI 53-58

Nam KH, Choe YC (2015) A study on onion wholesale price forecasting model. JAECD 22:423-434. doi:10.12653/jecd.2015.22.4.0423 10.12653/jecd.2015.22.4.0423

Olah C (2015) Understanding LSTM networks. https://colah.github.io/posts/2015-08-Understanding-LSTMs. 2021.03.17

Prakash N, Sumaiya Farzana G (2019) Short term price forecasting of horticulture crops using long short term memory neural network. Adv Comput Bio-Eng 111-118. doi:10.1007/978-3-030-46943-6_12 10.1007/978-3-030-46943-6_12

Shin SH, Lee MK, Song SK (2018) A prediction model for agricultural products price with LSTM network. Jour of KoCon a 18:416-429. doi:10.5392/JKCA.2018.18.11.416 10.5392/JKCA.2018.18.11.416

Information

Publisher :KOREAN SOCIETY FOR HORTICULTURAL SCIENCE
Publisher(Ko) :원예과학기술지
Journal Title :Horticultural Science and Technology
Journal Title(Ko) :원예과학기술지
Volume : 40
No :5
Pages :571-577
Received Date : 2022-02-03
Revised Date : 2022-04-07
Accepted Date : 2022-06-09
DOI :https://doi.org/10.7235/HORT.20220051

Horticultural Science and Technology ISSN:1226-8763(Print) 2465-8588(Online) 원예과학기술지

All Issue