Search results

The purpose of this paper is to study heterogeneous transfer learning for activity recognition using heuristic search techniques. Many pervasive computing applications require information about the activities currently being performed, but activity recognition algorithms typically require substantial amounts of labeled training data for each setting. One solution to this problem is to leverage transfer learning techniques to reuse available labeled data in new situations.

This paper introduces three novel heterogeneous transfer learning techniques that reverse the typical transfer model and map the target feature space to the source feature space and apply them to activity recognition in a smart apartment. This paper evaluates the techniques on data from 18 different smart apartments located in an assisted-care facility and compares the results against several baselines.

The three transfer learning techniques are all able to outperform the baseline comparisons in several situations. Furthermore, the techniques are successfully used in an ensemble approach to achieve even higher levels of accuracy.

The techniques in this paper represent a considerable step forward in heterogeneous transfer learning by removing the need to rely on instance – instance or feature – feature co-occurrence data.

The purpose of this paper is to determine and contrast the risk mitigating effectiveness from optimal multiproduct time-varying hedge ratios, applied to the margin of a cattle feedlot operation, over single commodity time-varying and naive hedge ratios.

A parsimonious regime-switching dynamic correlations (RSDC) model is estimated in two-stages, where the dynamic correlations among prices of numerous commodities vary proportionally between two different regimes/levels. This property simplifies estimation methods for a large number of parameters involved.

There is significant evidence that resulting simultaneous correlations among the prices (spot and futures) for each commodity attain different levels along the time-series. Second, for in and out-of-sample data there is a substantial reduction in the operation's margin variance provided from both multiproduct and single time-varying optimal hedge ratios over naive hedge ratios. Lastly, risk mitigation is attained at a lower cost given that average optimal multiproduct and single time-varying hedge ratios obtained for corn, feeder cattle and live cattle are significantly below the naive full hedge ratio.

The application studied is limited in that once a hedge position has been set at a particular period, it is not possible to modify or update at a subsequent period.

Agricultural producers, specifically cattle feeders, may profit from a tool using improved techniques to determine hedge ratios by considering a larger amount of up-to-date information. Moreover, these agents may apply hedge ratios significantly lower than one and thus mitigate risk at lower costs.

Feedlot operators will benefit from the potential implementation of this parsimonious RSDC model for their hedging operations, as it provides average optimal hedge ratios significantly lower than one and sizeable advantages in margin risk mitigation.