Taisir Eldos

Arabic Text Data Mining: a Root-Based Hierarchical Indexing Model

Abstract The world has recently witnessed a tremendous growth in the volume of text documents available on the Internet, digital libraries, news sources, and company-wide intranets. Text data mining, as a multidisciplinary field involving information retrieval, text analysis, information extraction, clustering, categorization, linguistics, database technology, machine learning, and data mining, is becoming more significant, and efforts have been intensified in studies like information retrieval, practical applications of which are becoming more and more necessary to end users and to the scientific community itself, in order to fetch the increasingly available information efficiently. In the past few years, not only have new documents been produced directly in digital form, thus being suitable for automatic indexing, but also many of the older documents have been ported from their physical medium to the digital one. The meaning of a document is represented by a vector of features, which are weighted according to a measure that best estimate relevance. Text categorization presents unique challenges due to the large number of attributes present in the data set, large number of training samples, and attributes dependencies. This article focuses on speeding up the information retrieval process in Arabic document base by using a root-based hierarchical indexing model. Simulation results demonstrated that speed gain in the range of 50-100 can be achieved for typical queries.

AN EFFICIENT CELL PLACEMENT USING GRAVITATIONAL SEARCH ALGORITHMS

In modern chip design, cell placement is a stage in which cells representing well-defined functions ar e assigned physical locations, in a way that optimize s the total area and routing length. Cell placement is an NP-complete problem and the exact solution is gener ally far from reach for a practically sized instanc e. Hence, diversified heuristic algorithms are used to solve this problem. In this study, we adapted a re cently introduced evolutionary search algorithm called Gra vitational Search Algorithm (GSA) to this problem. Experiments show that the proposed algorithm delivers good performance; good solution quality and likelihood of optimality within reasonably small am ount of time.

Performance Optimization of Adaptive Resonance Neural Networks Using Genetic Algorithms

We present a hybrid clustering system that is based on the adaptive resonance theory 1 (ART1) artificial neural network (ANN) with a genetic algorithm (GA) optimizer, to improve the ART1 ANN settings. As a case study, we will consider text clustering. The core of our experiments will be the quality of clustering, multi-dimensional domain space of ART1 design parameters has many possible combinations of values that yield high clustering quality. These design parameters are hard to estimate manually. We proposed GA to find some of these sets. Results show better clustering and simpler quality estimator when compared with the existing techniques. We call this algorithm genetically engineered parameters ART1 or ARTgep

A study of the robustness of the M-Max NLMS adaptive algorithm

In this paper, the M-Max NLMS algorithm is analyzed in a time-domain framework without constraining the number of updated coefficients per iteration or imposing statistical assumptions on the input signal. Analysis will prove the robustness of the M-Max NLMS algorithm under appropriate deterministic assumptions. Moreover, it will provide more insight into the algorithm structure leading to new choices of the step size for improved robust performance. A typical simulation example is presented to verify the deterministic findings.

Scalable autonic processing systems

The number of automotive functions that are controlled by computers is rapidly increasing. In the past, these functions were confined to stand-alone control units, such as the engine or ABS brake controller. In modern cars, the data exchange between different controllers has grown to a level where one or several networks are necessary to meet the communication demand [1]. Currently, automotive electronics processing is distributed in variety of functions like power train, air bag, climate control, suspension, etc. The purpose of this paper is to review related work through two state-of-the-art models, and propose a scalable design using parallel processing and fault tolerance techniques (excluding the functions of power train and transmission control electronics). The integration is based on the logical bisection of vehicle in four symmetric regions, which is critical for hardware, software, fault tolerance and ease of serviceability design. The proposed design can be implemented on any vehicle without going through the iterative system and module design cycles.

A functional memory based architecture for running sorting

In real-time applications, software-level implementations of running sorting algorithms may not be able to meet the processing time requirements, particularly when the size of the running window is large. In this paper, we present a hardware running sorter based on a functional memory architecture. The proposed approach accelerates operations thus giving the new hardware the capability of completing the running sorting algorithm in log N+7 CPU cycles. Details of the architecture units are explained, and a functional description of its operation provided.

Maximally Distant Codes Allocation Using Chemical Reaction Optimization and Ant Colony Optimization Algorithms

Error correcting codes, also known as error controlling codes, are set of codes with redundancy that allows detecting errors. This is quite useful in transmitting data over a noisy channel or when retrieving data from a storage with possible physical defects. The idea is to use a set of code words that are maximally distant from each other, hence reducing the chance of changing a valid codeword to another valid codeword by flipping bits. The problem can be viewed as picking m codes out of 2 n available n- bit combinations, such that the aggregate hamming distance among those codewords is maximized. Due to the large solution spaces of such problems, greedy algorithms are sometimes used to generate quick and dirty solutions. However, modern evolutionary search algorithms like genetic algorithms, swarm particles, gravitational search and others, offer good alternatives, yielding near optimal solutions in exchange for some time. Chemical Reaction Optimization (CRO) has emerged as a new evolutionary algorithm to solve complex optimization problems. This algorithm mimics the molecular interactions towards finding a minimal energy state, which corresponds to an optimal solution for the problem in hand. In this research, we proposed a solution for the maximally distant codes allocation problem, through a binary knapsack mapping and compared the performance with the well established Ant Colony Optimization (ACO) algorithm, which is inspired by the ants capability to find the shortest path between the nest and source of food. The binary knapsack mapping was used in the two algorithms. Test results showed that the CRO outperformed the ACO in every metric given any time budget.

A unified backward approach to the code book design problem

This paper presents a fast and robust approach for the code book design problem. Unlike the classical forward approach (CFA), which carries out the design process in two independent stages and in a forward direction; quantization then index assignment, the proposed unified backward approach (UBA) employs a statistics based mapping to generate the initial code vectors from the indices. This association faces a challenge due to the cardinally inequality, and this leads to utilizing the discrete cosine transform for compaction and dimension scaling. This mapping generates code vectors with two qualities; implicit tendency towards the constellation centroids and indices distance relationships analogy. Tests have shown that the UBA outperforms the CFA in more than 90% in terms of source distortion, while consistently having significant reduction in the channel distortion. Moreover, the improved initialization has caused some reduction in run time for higher bit rates designs.