Dumitru Iulian Nastac

Short-term financial forecasting using ANN adaptive predictors in cascade

Our purpose is to verify the predictive performances of the artificial neural networks (ANNs) under volatile statistics and possibly incomplete information. Daily forecasts of exchange rate using exclusively primary available information for an emergent economy (such as the Romanian one) could be a proper experimental ground with such a goal. The present paper extends the previous authors’ research (Dobrescu et al., 2006; Nastac et al., 2007) on the same issue to improve the accuracy of exchange rate forecasting by using a set of neural predictors in cascade, instead of a single one. The results show that the presented model, despite its proved advantages, could be further improved in order to avoid the translation into residuals of the high serial correlation present in the primary database.

A joint of adaptive predictors for electric load forecasting

This work describes a new approach of the adaptive retraining model for data forecasting. This time, six predictors are simultaneously employed in order to produce a better forecasting for electric load. By doing so, the new forecasting system eliminates iterative simulation. The set of predictors is regularly trained in order to be adjusted to the latest modifications of the input data. The new approach could be useful as a forecasting tool for a large variety of signals.

An advanced model for electric load forecasting

The paper describes an extension of the adaptive retraining model applied for electric load forecasting. Previously, the retraining procedure was successfully used for different kinds of applications that includes industrial, biomedical of financial data. This time the data comes from the most important power suppliers at the national level. Several cases are analyzed in order to find suitable parameters for a useful implementation.

Estimating profitability using a neural classification tool

The analysis of the private sector, mainly composed of micro, small and medium-sized business, briefly known as the SMEs sector, reveals many competitive aspects but also weaknesses, while the economic and social importance of the sector makes it imperative to elaborate specific development and consolidation strategies and policies. In order to establish such strategic directions, based on data on a significant volumes of Romanian SMEs (7902 entries) and on information provided by annual balance sheets, we established connections between results indicators (profit, measured as the rate of commercial profitability) and various causal variables by using a neural classification tool, which was built in accordance to testing and validation requirements in order to obtain consistent results.

A modified adaptive retraining procedure for data forecasting

The paper presents a further improvement of the adaptive retraining procedure of Artificial Neural Networks (ANNs) used for time series predictions. An important advantage of this approach is that the model is periodically adapted to the changes of the non-stationary environment. The retraining starts from proportionally reduced values of the parameters used in the previous version of the ANN model. As usual, variously delayed versions of the time series to be predicted and of the previous outputs are applied at the input of the ANN. In addition, the newly developed model also uses as inputs the averaged seasonal values from the previous years, obtained for the desired target variables in some specified time windows.

EMPHO-oxygradient I for monitoring in skin and skeletal muscle

This optical sensor system of high light sensitivity provides adequate prerequisites for tissue spectrometry in anesthesia and intensive care medicine. It can be applied for measurements of absolute hemoglobin concentration and oxygenation. In hemoglobin free perfused organs during transient hemoglobin free perfusion, monitoring of light scattering and oxygenation measurements of myoglobin and cytochromes are possible. Due to data base of spectra of organs high precision of optical parameters (microsecond(s) and (mu) a) can be attained. Calibration of light intensities of irradiated and backscattered light are performed by the computer system.

EMPHO-oxyscan spectrometer for 2D- and 3D-imaging in VIS and NIR

Oxyscan proved to be a very reliable optical sensor system for application in physiology and pathophysiology. Furthermore, it can be applied during liver transplantation in order to diminish or prevent reperfusion injury. Oxyscan could be used for monitoring of subcellular structures with great success.

EMPHO-oxygradient II for monitoring in organs during surgery and gastrointestinal diseases

Oxygradient II is an optical spectrometric sensor system for use in the microvolume of tissues (1-2 mm3). The microlightguides used for monitoring have catchment volumes of sizes that enable measurements of critical values at the venous end of capillaries. The Oxygradient II is able to be used for monitoring of all usual optical parameters in most human organs. A non-invasive monitoring during thoracic or abdominal surgery is possible with less-invasive micro-lightguides of diameters of 600 micrometers in combination with thin endoscopes.

Quantitative monitoring of drug-tissue interaction in perfused organs

In pharmacology many optical sensors are applied for investigations in vitro and in reduced systems. Due to a lack of sensors for optical imaging of functional structures in capillaries as well as in subcellular spaces the drug-tissue interaction in organs could not be monitored systematically. However, recent developments opened the door of this microcosm of life in its smallest entities. This will enable a better understanding of the questions of area and quality of drug action in tissue.