Guidance on the choice of threshold for binary forecast modeling

Sohn, Keon Tae; Park, Sun Min

doi:10.1007/s00376-008-0083-8

Cited by 12 publications

(8 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…One obvious choice ω t is 0.5, but other values could also be used. In particular, the threshold value ω t is often chosen in such a way that some skill score is maximized (Sohn & Park 2008). However, as the optimal threshold ω t depends on the chosen skill score, it would also depend on the corresponding economic value function or objective function.…”

Section: A Test Of the Probability Forecastsmentioning

confidence: 99%

Testing the value of probability forecasts for calibrated combining

Lahiri

Peng

Zhao

2015

International Journal of Forecasting

View full text Add to dashboard Cite

We combine the probability forecasts of a real GDP decline from the U.S. Survey of Professional Forecasters, after trimming the forecasts that do not have “value”, as measured by the Kuiper Skill Score and in the sense of Merton (1981). For this purpose, we use a simple test to evaluate the probability forecasts. The proposed test does not require the probabilities to be converted to binary forecasts before testing, and it accommodates serial correlation and skewness in the forecasts. We find that the number of forecasters making valuable forecasts decreases sharply as the horizon increases. The beta-transformed linear pool combination scheme, based on the valuable individual forecasts, is shown to outperform the simple average for all horizons on a number of performance measures, including calibration and sharpness. The test helps to identify the good forecasters ex ante, and therefore contributes to the accuracy of the combined forecasts.

show abstract

Section: A Test Of the Probability Forecastsmentioning

confidence: 99%

Testing the value of probability forecasts for calibrated combining

Lahiri

Peng

Zhao

2015

International Journal of Forecasting

View full text Add to dashboard Cite

show abstract

“…The SPI is presented at 3, 6, 12, 24 and 48 months time steps that start in October, except for the January to March sub-season. Validation of the binary forecast model uses a 2 × 2 contingency table (Table 4) and a range of skill scores as in Sohn and Park (2008).…”

Section: Validation Of Findingsmentioning

confidence: 99%

“…Probabilistic information is difficult to assimilate because farmers do not think probabilistically nor do they interpret probabilities easily for decisionmaking (Nicholls, 1999). It has also been argued that, the probabilistic forecast cannot be used directly when estimated too smoothly (Sohn & Park, 2008). In such cases, the binary forecast is preferable to the probabilistic forecast.…”

Section: Introductionmentioning

confidence: 99%

Tailoring seasonal climate forecasts for climate risk management in rainfed farming systems of southeast Zimbabwe

Unganai¹,

Troni²,

Manatsa

et al. 2013

Climate and Development

View full text Add to dashboard Cite

Inter-annual climate variability is a major source of production risks in rainfed agriculture in semi-arid areas of Zimbabwe. Despite advances in seasonal forecasting since the late 1990s, there is hardly any evidence of explicit use of forecasts by vulnerable smallholder farmers to manage climate-related risk in agriculture. Forecasts are presented in the language of probabilities, but are often not perceived as such, partly because of differences in end users' needs and their decisionmaking behaviour. The project used the Indian Ocean Dipole (IOD) index and El Niño-Southern Oscillation (ENSO) to develop a tailored seasonal forecast model termed the 'binary' or 'drought/no drought'. The binary forecast model addresses farmers most important question: what is the probability of drought (SPI ≤ 21) occurring in the crop growing season? The 'binary' forecast system allowed the development of climate risk management strategies specifically tailored to farmers' needs. From this study, it can be concluded that rainfed agriculture production systems are most concerned about the risk of drought. A tailored forecast system that provides information on the probability of drought occurring in a given season can therefore lead to proactive drought risk management among smallholder farmers and policy-makers.

show abstract

“…Sohn 등 (2009)은 호남지역 대설특보를 위하여 로지스틱 회귀모형과 신경회로망을 적용하였다. Sohn과 Park (2008)은 이 범주 예 보를 위한 문턱치 결정을 위하여 예측성 평가측도를 활용하는 가이던스를 제안하였다. 로지스틱 회귀모 형은 순서형 자료를 예측량으로 하는 확률예측모형으로 유용하게 사용된다.…”

unclassified

Binary Forecast of Asian Dust Days over South Korea in the Winter Season

Sohn¹,

Kim²

2011

Korean Journal of Applied Statistics

View full text Add to dashboard Cite

Guidance on the choice of threshold for binary forecast modeling

Cited by 12 publications

References 8 publications

Testing the value of probability forecasts for calibrated combining

Testing the value of probability forecasts for calibrated combining

Tailoring seasonal climate forecasts for climate risk management in rainfed farming systems of southeast Zimbabwe

Binary Forecast of Asian Dust Days over South Korea in the Winter Season

Contact Info

Product

Resources

About