انجمن کامپیوتر ایران

در سيستم هاي شناسايي زبان كه از اطلاعات سطح پايين زباني استفاده مي كنند، معمولا از يك پس پردازنده براي جداسازي بهتر زبان ها از هم استفاده مي شود. پس پردازنده هاي رايج از نوع مدل هاي مخلوط گوسي مي باشند كه نياز به دادگان وسيع توسعه اي دارند. از طرفي تصميم گيري روي بردارهاي خروجي هر فريم گفتاري براي شناسايي زبان، دقت را بسيار پايين مي آورد. براي رفع اين مشكل ابتدا توسط كرنل هاي دنباله اي، كل بردارهاي حاصل از يك گفتار را به يك بردار نگاشت مي كنند و سپس توسط ماشين هاي بردار پشتيبان، اين بردارها را از هم جدا مي كنند. ما در اين مقاله كرنل دنباله اي جديدي پيشنهاد كرده ايم كه نسبت به كرنل هاي دنباله اي مشابه كه بهترين راندمان را در كاربرد شناسايي زبان دارند و همچنين پس پردازنده هاي گوسي، راندمان سيستم شناسايي زبان را بالاتر مي برد. ما آزمايش هاي خود را روي دو دادگان 2003& 1996 NIST-LRE و OGI-TS 11و 22 زباني انجام داده ايم. نتايج نشان دهنده افزايش نسبي حدودا 4% راندمان شناسايي زبان با استفاده از كرنل دنباله اي پيشنهادي ما مي باشند.

كاوش داده هايي مانند ميكروآرايه ها كه ابعاد بس يار بالا يي دارند، باعث به وجود آمدن مشكلات جد ي بر اي اكثر الگور يتم ه اي كارآمد الگوكاوي موجود مي گردد و كارآ يي آنها را با چالش جد ي مواجه مي نمايد. در سالهاي اخير متدهاي جديدي ارائه شده كه مبتني بر شمارش سطرها هستند و با انجام يك جستجوي از پايين به بالا بر روي فضاي تركيب سطرها سعي در الگوكاوي كارآمد دارند كه با توجه به تعداد اندك سطرها در داده هاي ميكروآرايه اي، اين روش ها بس يار كارآمدتر از روش هاي مبتني بر شمارش ستون ها هستند. اما رويكرد از پايين به بالاي اين روش ها از يك سو و عدم امكان فشرده سازي مناسب اطلاعات از سوي ديگر سبب شده است كه روش هاي مزبور فاقد كارآيي كافي به منظور هرس نمودن فضاي جس تجو با توجه به مقدار حد آستانه مشخص شده توسط كاربر باشند و اين مساله منجر به طولاني شدن زمان اجرا و ايجاد فضاي سربار در حافظه مي گردد. در اين مقاله روش جستجوي جديدي به صورت كاوش از بالابه پا يين مبتني بر نمايش بيتي تراكنش ها بر روي ي ك درخت شمارش سطر ارائه مي گردد كه امكان استفاده كامل از قدرت هرس كردن توسط حد آستانه اي داده شده توسط كاربر را فراهم مي نمايد و به كوچك شدن شگفت انگيز فضاي جستجو مي انجامد . با استفاده از ا ين استراتژ ي جستجو، الگوريتم كاوشي به نام Clobimat طراحي شده است كه قادر به يافتن مجموعه كامل همه الگوه اي تكرارشونده بسته در مجموعه داده هاي با ابعاد بسيار بالاست. به علاوه روش جديدي براي تست بسته بودن الگوها ارائه شده است كه مانع از پويش چندين باره مجموعه داده مي گردد. مطالعات انجام شده نشان داده است كه الگوريتم ارائه شده در اين مقاله به لحاظ كارآيي برتري قابل ملاحظه اي نسبت به الگور يتم Carpenter كه يك الگور يتم جستجوي پايين به بالا ي فضاي سطرهاست- و الگوريتم FPClose كه يك الگوريتم مبتني بر جستجوي ستون هاست- دارد.

In recent years many researches have focused on ad-hoc networks, mainly because of their independence to any specific structure. These networks suffers from frequent and rapid topology changes that cause many challenges in their routing. Most of the routing protocols try to find a path between source and destination nodes because any path will expire, offer a short period, the path reconstruction may cause the network inefficiency. The proposed protocol build two paths between source and destination and create backup paths during the route reply process, route maintenance process and local recovery process in order to improve the data transfer and the fault tolerance. The protocol performance is demonstrated by using the simulation results obtain from the global mobile simulation software(Glomosim). The experimental results show that this protocol can decrease the packet loss ratio rather than DSR and SMR and it is useful for the applications that need a high level of reliability.

The traffic congestion problem in urban areas is worsening since traditional traffic signal control systems cannot provide efficient traffic control. Therefore, dynamic traffic signal control in Intelligent Transportation System (ITS) recently has received increasing attention. This study devised an adaptive and cooperative multi-agent fuzzy system for a decentralized traffic signal control. To achieve this goal we have worked on a model, which has three levels of control. Every intersection is controlled by its own traffic situation, its neighboring intersections recommendations and a knowledge base, which provides the traffic pattern of each intersection in any particular day of the week and hour of the day. The proposed architecture comprises a knowledge base, prediction module and a traffic observer that provide data to real traffic data preparation module, then a decision-making layer takes decision to how long should the intersection green light be extended. The proposed architecture can achieve dynamic traffic signal control. We have also developed a NetLogobased traffic simulator to serve as the agents’ world. Our approach is tested with traffic control of a large connected junction and the result obtained is promising; The average delay time can be reduced by 21.76% compared to the conventional fixed sequence traffic signal and 14.77% compared to the vehicle actuated traffic signal control strategy.

One of the important requirements for operational planning of electrical utilities is the prediction of hourly load up to several days, known as Short Term Load Forecasting (STLF). Considering the effect of its accuracy on system security and also economical aspects, there is an on-going attention toward putting new approaches to the task. Recently, Neuro Fuzzy modeling has played a successful role in various applications over nonlinear time series prediction. This paper presents a neuro-fuzzy model for the application of short-term load forecasting. This model is identified through Locally Liner Model Tree (LoLiMoT) learning algorithm. The model is compared to a multilayer perceptron and Kohonen Classification and Intervention Analysis. The models are trained and assessed on load data extracted from EUNITE network competition.

Portfolio optimization under classic mean-variance framework of Markowitz must be revised as variance fails to be a good risk measure. This is especially true when the asset returns are not normal. In this paper, we utilize Value at Risk (VaR) as the risk measure and Historical Simulation (HS) is used to obtain an acceptable estimate of the VaR. Also, a well known multi-objective evolutionary approach is used to address the inherent bi-objective problem; In fact, NSGA-II is incorporated here. This method is tested on a set of past return data of 12 assets on Tehran Stock Exchange (TSE). A comparison of the obtained results, shows that the proposed method offers high quality solutions and a wide range of risk return trade-offs.

A watermarking method in DCT domain is modified to achieve better imperceptibility. Particle Swarm Optimization (PSO) is used to find the best DCT coefficients for embedding the watermark sequence and the Structural Similarity Index is used as the fitness function in order to have a watermarked image with the best possible quality.

A model of episodic memory is derived to propose algorithms of text categorization with semantic space models. Performances of two algorithms are contrasted using textual material of the text-mining context ‘DEFT09’. Results confirm that the episodic memory metaphor provides a convenient framework to propose efficient algorithm for text categorization. One algorithm has already been tested with LSA. The present paper extends these algorithms to another model of Word Vector named Random Indexing.