• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권5호
• /
• pp.1414-1430
• /
• 2022

Due to the Internet of Things popularity, many agricultural data are collected by sensors automatically. The abundance of agricultural data makes precise prediction of rice yield possible. Because the climate factors have an essential effect on the rice yield, we considered the climate factors in the prediction model. Accordingly, this paper proposes a machine learning model for rice yield prediction in Taiwan, including the genetic algorithm and support vector regression model. The dataset of this study includes the meteorological data from the Central Weather Bureau and rice yield of Taiwan from 2003 to 2019. The experimental results show the performance of the proposed model is nearly 30% better than MARS, RF, ANN, and SVR models. The most important climate factors affecting the rice yield are the total sunshine hours, the number of rainfall days, and the temperature.The proposed model also offers three advantages: (a) the proposed model can be used in different geographical regions with high prediction accuracies; (b) the proposed model has a high explanatory ability because it could select the important climate factors which affect rice yield; (c) the proposed model is more suitable for predicting rice yield because it provides higher reliability and stability for predicting. The proposed model can assist the government in making sustainable agricultural policies.

• 대한산업공학회지
• /
• 제41권6호
• /
• pp.572-578
• /
• 2015

Yield prediction is one of the most important issues in semiconductor manufacturing. Especially, for a fast-changing environment of the semiconductor industry, accurate and reliable prediction techniques are required. In this study, we propose a prediction model to predict wafer yield based on virtual metrology process parameters in semiconductor manufacturing. The proposed prediction model addresses imbalance problems frequently encountered in semiconductor processes so as to construct reliable prediction model. The effectiveness and applicability of the proposed procedure was demonstrated through a real data from a leading semiconductor industry in South Korea.

• 대한산업공학회지
• /
• 제21권3호
• /
• pp.371-385
• /
• 1995

The yield of semiconductor chips is dependent not only on the average defect density but also on the distribution of defects over a wafer. The distribution of defects leads to consider a cluster index. This paper briefly reviews the existing yield prediction models ad proposes a new cluster index, which utilizes the information about the defect location on a wafer in terms of the coefficient of variation. An extensive simulation is performed under a variety of defect distributions and a yield prediction model is derived through the regression analysis to relate the yield with the proposed cluster index and the average number of defects per chip. The performance of the proposed simulation-based yield prediction model is compared with that of the well-known negative binomial model.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제16권3호
• /
• pp.157-162
• /
• 2016

In this paper, a yield prediction model based on relevance vector machine (RVM) and a granular computing model (quotient space theory) is presented. With a granular computing model, massive and complex meteorological data can be analyzed at different layers of different grain sizes, and new meteorological feature data sets can be formed in this way. In order to forecast the crop yield, a grey model is introduced to label the training sample data sets, which also can be used for computing the tendency yield. An RVM algorithm is introduced as the classification model for meteorological data mining. Experiments on data sets from the real world using this model show an advantage in terms of yield prediction compared with other models.

• 한국농공학회논문집
• /
• 제54권3호
• /
• pp.141-148
• /
• 2012

Food policy is considered as the most basic and central issue for all countries, while making efforts to keep each country's food sovereignty and enhance food self-sufficiency. In the case of Korea where the staple food is rice, the rice yield prediction is regarded as a very important task to cope with unstable food supply at a national level. In this study, Korean paddy Rice yield Prediction Model (KRPM) developed to predict the paddy rice yield using meteorological element and MODIS NDVI. A multiple linear regression analysis was carried out by using the NDVI extracted from satellite image. Six meteorological elements include average temperature; maximum temperature; minimum temperature; rainfall; accumulated rainfall and duration of sunshine. Concerning the evaluation for the applicability of the KRPM, the accuracy assessment was carried out through correlation analysis between predicted and provided data by the National Statistical Office of paddy rice yield in 2011. The 2011 predicted yield of paddy rice by KRPM was 505 kg/10a at whole country level and 487 kg/10a by agroclimatic zones using stepwise regression while the predicted value by KOrea Statistical Information Service was 532 kg/10a. The characteristics of changes in paddy rice yield according to NDVI and other meteorological elements were well reflected by the KRPM.

• 한국환경과학회지
• /
• 제18권1호
• /
• pp.1-7
• /
• 2009

In this study a sediment yield is compared by IUSG, IUSG with Kalman filter, tank model and tank model with Kalman filter separately. The IUSG is the distribution of sediment from an instantaneous burst of rainfall producing one unit of runoff. The IUSG, defined as a product of the sediment concentration distribution (SCD) and the instantaneous unit hydrograph (IUH), is known to depend on the characteristics of the effective rainfall. In the IUSG with Kalman filter, the state vector of the watershed sediment yield system is constituted by the IUSG. The initial values of the state vector are assumed as the average of the IUSG values and the initial sediment yield estimated from the average IUSG. A tank model consisting of three tanks was developed for prediction of sediment yield. The sediment yield of each tank was computed by multiplying the total sediment yield by the sediment yield coefficients; the yield was obtained by the product of the runoff of each tank and the sediment concentration in the tank. A tank model with Kalman filter is developed for prediction of sediment yield. The state vector of the system model represents the parameters of the tank model. The initial values of the state vector were estimated by trial and error.

• Agricultural and Biosystems Engineering
• /
• 제3권1호
• /
• pp.18-28
• /
• 2002

Neural network based models were developed and evaluated for predicting corn yield from aerial images based on 1998 and 1994 image data. The model used images in multi-spectral bands such as R, G, B, and IR (Red, Green, Blue and Infrared). The inputs to the neural network consisted of mean and standard deviation of multispectral bands of the aerial images. Performances of several neural network architectures using back-propagation with momentum were compared. The maximum yield prediction accuracy obtained was 97.81%. The BPNN model prediction accuracy could be enhanced by using more number of observations to the model, other data transformation techniques, or by performing optical calibration of the aerial image.

• Asian-Australasian Journal of Animal Sciences
• /
• 제11권5호
• /
• pp.580-585
• /
• 1998

The study was conducted to determine if variation in protein yield can be explained by expressions of early lactation somatic cell score (SCS) and if prediction can be improved by including SCS among the predictors. A data set was prepared (n = 663,438) from Wisconsin Dairy Improvement Association (USA) records for protein yield with sample days near 20. Stepwise regression was used requiring F statistic (p < .01) for any variable to stay in the model. Separate analyses were run for 12 combinations of four seasons and first three parities. Selection of SCS variables was not consistent across seasons or lactations. Coefficients of detennination ($R^2$) ranged from 51 to 61% with higher values for earlier lactations. Including any expression of SCS in the prediction equations improved $R^2$ by < 1 %. SCS was associated with milk yield on the sample day, but the association was not strong enough to improve the prediction of future yield when other expressions of milk yield were in the model.

• 한국농림기상학회지
• /
• 제21권2호
• /
• pp.75-84
• /
• 2019

이 연구의 목적은 random forest 를 활용하여 기상요소만을 이용하여 우리나라 전체의 벼 평균수량을 예측하는데 있다. Random forest 는 예측에 사용되는 각 predictor variable 을 분리할 수 있는데 이를 통해 분리된 시계열 상의 추세가 비정상적인 증가형태를 보였다. 이는 결국 예측능력의 저하로 이어지기 때문에 이를 제거할 필요가 있고 본 연구에서는 이동 평균을 이용하여 제거한 후 예측을 하였다. 1991 년부터 2005 년까지의 기상자료와 수량자료를 학습에 사용하였고 2006 년부터 2015 년까지의 자료들을 검증용으로 사용하였다. 학습자료에 대해서는 상당히 정확한 예측 능력을 보여주었으나 검증 자료에서는 그렇지 못하였다. 그 이유를 분석하기 위해 학습 자료와 검증자료에 대해서 각각 변수 중요도를 산출하여 비교한 결과 두 자료 간에 월별 기상 자료에 대한 중요도가 변동되었음을 발견하였다. 이러하 차이가 발생한 이유는 학습자료와 검증 자료에서의 전국적으로 표준이앙기가 이동하여 벼의 생육기간 자체가 변하였기 때문이다. 따라서, 정확한 예측을 위해서는 지역별 파종기 또는 이앙기에 대한 자료가 필요하며 단순히 기상 자료만을 활용한 예측은 어려운 것으로 생긱된다.

• 한국농림기상학회지
• /
• 제4권3호
• /
• pp.164-168
• /
• 2002

우리나라 벼 수량 기상반응의 최대경계선 분석을 통하여 구축한 벼 수량예측모델(금 등, 2001)의 지역, 연차 및 품종의 수량 변이 예측 정확도를 검증한 결과는 다음과 같다. 1. 모형구축에 이용된 전국 20개 지역의 지역별 15년 간의 평균수량은 실측치와 예측치 간에 r=0.9296$^{**}$ 으로 고도로 유의한 상관을 나타내고 있으며 모형작성에서 제외시켰던 12개 지역 평균수량의 예측치와 실측치간의 상관계수도 r=0.8923$^{**}$ 으로 모형작성에 이용된 지역보다 수량예측의 정확도가 다소 낮게 나타났으나 통계적으로 고도로 유의하였다. 2. 모형설정에 사용된 20개 지역과 사용되지 않는 12개 지역 모두 연차별 실제수량과 예측수량간에는 고도로 유의한 상관이 존재하여 수량의 연차간 변이를 잘 예측하였다. 다만 냉해년에는 다소과대 추정하는 경향이었다. 3. 동진벼, 화성벼, 추청벼 등의 8개 품종별로 실제 수량과 예측수량간에는 고도로 유의한 상관이 있었다. 다만 모형구축에 이용되지 않은 지역의 자료를 이용하는 경우 모형설정에 이용한 자료를 이용한 경우보다 다소 상관이 낮아졌다. 4. 결론적으로 수량기상반응의 최대경계선(boundary line)분석은 수량 예측 모형의 구축에 효율적으로 적용될 수 있을 것으로 판단되었으며, 본 연구에서 고려하지 않은 토양조건, 시비조건 등에 대한 최대경계선 분석을 포함시키는 경우 보다 정확도가 높은 수량예측모형을 작성할 수 있으며 이를 통한 지성, 년도, 품종에 따른 수량의 변이를 실용적으로 예측할 수 있을 것으로 기대된다.