Manana Khachidze

Natural Language Processing Based Instrument for Classification of Free Text Medical Records.

According to the Ministry of Labor, Health and Social Affairs of Georgia a new health management system has to be introduced in the nearest future. In this context arises the problem of structuring and classifying documents containing all the history of medical services provided. The present work introduces the instrument for classification of medical records based on the Georgian language. It is the first attempt of such classification of the Georgian language based medical records. On the whole 24.855 examination records have been studied. The documents were classified into three main groups (ultrasonography, endoscopy, and X-ray) and 13 subgroups using two well-known methods: Support Vector Machine (SVM) and K-Nearest Neighbor (KNN). The results obtained demonstrated that both machine learning methods performed successfully, with a little supremacy of SVM. In the process of classification a “shrink” method, based on features selection, was introduced and applied. At the first stage of classification the results of the “shrink” case were better; however, on the second stage of classification into subclasses 23% of all documents could not be linked to only one definite individual subclass (liver or binary system) due to common features characterizing these subclasses. The overall results of the study were successful.

The method of concept formation for semantic search

The article considers the method of forming concepts forming can be used in semantic search of information. The said method has been worked out by way of synthesizing two existing methods: Explicit Semantic Analysis (ESA) and analytical heuristics. The article examines a new and more convenient method for describing a concept (its attributes) and for calculating concepts. This method describes concepts more compactly and semantically more precisely. The described method together with other methods can be used in conceptual search.

Some Size and Quantum Effects in Molecular Nanostructures

Quantum effects in nanostructured magnetic solid state materials open the new ways for preparing the novel electromagnetic devices with unique characteristics. At the same the electron spin based quantum effects are fully determined of formation and growing properties of molecular clusters of these solid state materials which are united in plane and volume structures also according their quantum (elementary particle interactions) properties.

Quantum Approach to Sensory Information Processing for Modeling of Disasters

In the modeling process of big systems such as different disasters, one of the main points is data collection: data quantity and quality, its manifold and accuracy. The modeling of this kind of processes includes: data mining and recognition (creation of a database), data processing and treatment (knowledge base preparation), and elaboration of conclusions. The quantum information technologies permit by using quantum representation of data, to collect much bigger, more varied and precise information, as well as quantum databank creation, which can be effectively treated by usage of relevant quantum algorithms. For creation of quantum databases, two methods will be dealt with: One is based on quantum numbers usage for processing various parametrical values (attributes of database); next, the database will be presented as its quantum model. Taking into account the quantum and multi parametrical nature of disasters for their clear and precise modeling it is necessary to combine the above mentioned two approaches and jointly quantum search algorithms and quantum query algorithms as well.

Complex system state generalized presentation based on concepts

The new generation technologies such as nanosensors, are allowing the monitoring and evaluation of complex systems with a microscopic nodes. Nowadays it is possible to have nanosensors in such type of nodes and it will not worsen the quality of their work. In the proposed work we will be dealing with system state automatic evaluation problem and try to develop the automatic evaluation expert system knowledgebase representation mechanism for system state, based on concepts pattern.

Quantum concepts in information retrieval

Quantum information methods are successfully used with classic methods in information retrieval. This paper reviews the formation of quantum concepts, their presentation and storage and comparison methods for algorithms of information retrieval. We survey the representation of quantum concepts with qubits and their comparison methods via quantum logic gates. The paper discusses fuzzy concepts ranking, relevance evaluation and machine learning methods to improve information retrieval.

Photoelectric characteristics of contacts In-semiconductor A3B5

The effect of annealing structures on the electrical and photoelectric properties of metal-semiconductor contacts was investigated. Metal/semiconductor structures have been fabricated by method of electrochemically deposition of In on the electrochemically cleaned surface of the semiconductors A3B5 (GaP, GaAs). The dark capacitance and current -voltage characteristics and the hotoelectric spectra of zero bias for front-illuminated contact show near-ideal Schottky barrier diode properties for annealing temperature up to 250-3000C. Was found that the spectra of zero bias photocurrent of In/GaP beside the region photoconductivity resulting from band to band excitation, contains, also, separated of them the region photoconductivity in a long wavelength of spectra, which is related to the interaction between the metal and semiconductor. Samples used for the fabrication of In/GaP diodes were growing by Chochralski method especially un doped n-type GaP into (III) oriented wafers. The thickness and carrier concentration was 200-250 mimic and (2-4). 10 exp17 atom/cm3 respectively. At first ohmical contact to the one side of wafer was formed by alloying of indium at the temperature 5000C for GaAs and 600°C for GaP during 5 min in hydrogen. Then the sample with ohmic contact and wire for preceding the power was coaled with chemical stable polystyrene solution except the area where the metal will be deposited. The wafers were then ached chemically, rinsed in distilled water and were transferred immediately into electrolyte for deposition of In.

Electrochemical deposition of Al on semiconductors

Samples used for the fabrication of M-S diodes were growing by Chochralski method especially undoped n-type GaP into (III) oriented wafers. The thickness and carrier concentration was 200-250 mimic and (2-4). 10 exp17 atom/cm3 respectively. At first ohmical contact to the one side of wafer was formed by alloying of indium at the temperature 600°C during 5 min in hydrogen. Then the sample with ohmical contact and wire for preceding the power was covered with chemical stable polystyrene solution except the area where the metal will be deposited. The wafers were then ached chemically, rinsed in distilled water and were transferred immediately into electrolyte. Deposition of metal was done by the usual electrochemical method. Electrolyte was poured into quartz glass. The semiconductors wafer was used as the one electrode and as another electrode was used aluminum. For deposition Al the aqueous solution of chlorides have been used as an electrolyte, which consisted also NaOCI. At first, semiconductor’s wafer was used as the anode and cleaning of semiconductors surface was done. Then the potential was immediately changed in opposite direction and deposition of metal on freshly cleaned surface was done in the same solution in a united technological process. After the process of realization the samples were washed in distilled water. The polystyrene film was removed mechanically and boiling in acetone. Then samples were cut into pieces of area 1-3 mm, and were measured electric and photoelectric characteristics. The electrical and photoelectric characteristics have been studded and they were analyzed in the usual way to calculate the ideality factor (n) and barrier height (o). The values of coefficient n and SB height were 1.05-1.07 and 1.1 eV respectively.