Classification Cramer-Rao bounds on stock price prediction
Article Abstract:
An innovative approach to quantifying the classification Cramer-Rao lower bound, which involves a performance metric for a classifier given input features, is developed. It combines the concepts of sufficient statistics and data compression with a metric that measures class discrimination. The approach allows the assessment of the degree of performance optimality attained by each classifier as it assigns a quantity called the optimality score that indicates the extent of a classifier's approximation of the Bayes classifier. Its potency is demonstrated with two-class prediction problems.
Publication Name: Journal of Forecasting
Subject: Mathematics
ISSN: 0277-6693
Year: 1998
User Contributions:
Comment about this article or add new information about this topic:
Judgemental revision of sales forecasts: the relative performance of judgemntally revised versus non-revised forecasts
Article Abstract:
The increasingly common practice of judgmental revision of sales forecasts is analyzed for an industrial company through six quarters and some 900 products. The error distributions in forecasts after managers' revisions are compared to those without revision, using Kolmogorov-Smirnov, medians and t tests. However, it is concluded that non-revised forecasts generally have lesser errors than revised ones and that harmonization between the two types is the best result one can achieve. Revision also improves forecast accuracy but is insignificant where variability is concerned.
Publication Name: Journal of Forecasting
Subject: Mathematics
ISSN: 0277-6693
Year: 1992
User Contributions:
Comment about this article or add new information about this topic:
Prediction in the one-way error component model with serial correlation
Article Abstract:
The best linear unbiased predictor is formulated for a one-way error component model with serial correlation. It can be easily computed from the GLS estimates and residuals through the use of a transformation, and is applicable to panel data cases which utilize the error component specification and exhibit serial correlation in the remainder disturbance term. This model is then analyzed wherein the remainder disturbances follow either an AR(1), an AR(2), an MA(1) or a special AR(4) process for quarterly data.
Publication Name: Journal of Forecasting
Subject: Mathematics
ISSN: 0277-6693
Year: 1992
User Contributions:
Comment about this article or add new information about this topic:
- Abstracts: The impact of measurement errors on ARMA prediction. Organizational pressures on forecast evaluation: managerial, political, and procedural influences
- Abstracts: Estimating deterministic trends in the presence of serially correlated errors. Money, prices, interest rates and the business cycle
- Abstracts: Variance estimation for multivariate dynamic linear models. Bias in the estimation of non-linear transformation of the integrated variance of returns
- Abstracts: The dynamic and stochastic instability of betas: implications for forecasting stock returns. Forecasting interest rates and yield spreads: the informational content of implied futures yields and best-fitting forward rate models