61 research outputs found
Naive Bayes classifier for word sense disambiguation of Punjabi language
No full textWord Sense Disambiguation (WSD) is the process of identifying the correct sense of the word in the context. The most leading scheme used by WSD is machine learning approach, where a human expert provides examples of correctly disambiguated words, and a machine learning algorithm is used to induce a model from these examples. In this paper, Naive Bayes supervised classifier has been used to disambiguate words of Punjabi language. The feature extraction process plays a vital role in building the supervised machine learning models. For the proposed Punjabi WSD system, Bag of Words (BoW) and collocation models are used separately to extract relevant features. BoW model has used all words around target word while collocation model has used two words before and two words after the target word as features. Both the models have used a common training data set to build the model. It has been observed that the selection of smoothing parameter for Naive Bayes has a significant impact on its performance. This proposed work has been tested on 150 most ambiguous noun words selected form Punjabi WordNet having 6 or more senses. During the process of building the model, fine senses of ambiguous words have been merged to produce coarse sense on the basis of manual analysis of lexical relations of WordNet. The accuracy of the proposed system has been calculated independently for BoW and collocation model. The proposed WSD system achieves an accuracy of 89% for BoW model and 81% for collocation model. It has been concluded that BoW model performs better than the collocation model for WSD task for Punjabi language
Design and Implementation of Computer and Network Forensics Framework
No full textDoctor of Philosophy -CSEWith an exponential increase in the data size and complexity of various seized items to be investigated, existing methods of network and computer forensics are not very efficient when it comes to dealing with accuracy and detection ratio. Till the time a well-established forensic technique is developed to handle security threats, a much more sophisticated attacks strike on network. Traditional Intrusion Detection Systems (IDS) and forensics techniques used to detect and prevent malicious network behaviours, fail to handle new or zero day attacks. The accuracy of Intrusion Detection Systems (IDS) and Intrusion Prevention Systems (IPS) is questionable, which can’t be trusted for forensics. Another important drawback with the exiting techniques, is their inability to tackle high velocity and huge amount of heterogeneous data.
Cyber forensic investigation mechanism has volume constraint, while processing the fast growing data from Information and Communication Technology (ICT) infrastructure, including IoT based devices and platforms. Non-tangible sources often don’t have the limit of flowing data through them, especially through communication media. Hence, increasing the desperate requirement for an efficient benchmarking of big data analysis. Existing techniques exhibit inherent limitations in processing huge volume, variety, and velocity of data. It makes the process time-consuming and resource intensive. Available solutions to date have used an anomaly-based approach or have proposed approaches based on the deviation from a regular pattern. To tackle the seized bytes, authors have proposed an approach for big data forensics, with efficient sensitivity and precision.
In order to maintain a balance between processing time and output efficiency, existing techniques put a limit on the amount of data under analysis, which results in a non-polynomial time complexity of these solutions. In this thesis, a scalable, practical framework to overcome the limitation to handle large volume, variety, and velocity of data, is proposed. The proposed architectural setup consists of the MapReduce framework on top of the Hadoop Distributed File System environment.
The proposed framework demonstrates its capability to handle issues of storage and processing of big data using cloud computing infrastructure. In the presented work, a
generalized forensic framework has been proposed that use Google’s programming model, MapReduce as the backbone for traffic translation, extraction, and analysis of dynamic traffic features. For the proposed technique, authors have used open source tools like Hadoop, Hive, and Mahout and R. Apart from being open source, these tools support scalability and parallel processing. Also, comparative analysis of globally accepted machine learning models of P2P malware analysis in mocked real-time is presented. Supervised machine learning (Random Forest based Decision Tree) algorithm has been implemented to demonstrate better sensitivity and specificity. For training and validating the model, CAIDA dataset [1] along with university network traffic samples from GitHub [2], with increasing size, has been taken. Results thus obtained confirm the superiority of the proposed framework, with an accuracy of 99%. The work encompasses computer and network forensics, which is being referred as cyber forensics, collectively in this thesis, due to the nature of the data being dealt and experimented
Retrofitting of reinforced concrete beam-column joints using bonded laminates
No full textPh.D. (Civil Engineering)Reinforced concrete is one the most abundantly used construction material, not only in the developed world, but also in remotest parts of the developing world. Thousands of reinforced concrete structures are constructed annually and a large number of these deteriorate or become unsafe before the end of their design life. Strengthening of existing reinforced concrete structures is now a major component of construction activity. The RCC structures constructed across the world are often found to exhibit distress and suffer damage, even before service life is over, due to several causes such as earthquakes, corrosion, overloading, improper design, faulty construction, explosions, fire etc. With the mandate to go vertical, due to rising population and space crunch, most of the structures which have come up over the last three or more decades are all framed structures. In such structures, the most important link for transferring loads and stresses are the beam-column joints. The structural design of these joints is usually neglected. Unsafe designs and detailing within the joint region is dangerous for the entire structure, even though the structural members themselves may conform to the design requirements. It is well known that the joint regions in reinforced concrete framed structures are very critical as they transfer the forces and bending moments between the beams and columns. In most cases, during extreme loading, the beam-column joints, if not designed properly are the most vulnerable component. With the advent of revised design and detailing codes and increase in the earthquake vulnerability level of many regions, the existing structures need strengthening to conform to the revised codal provision.
The strengthening and enhancement of the performance of deficient structural elements in a structure or the structure as a whole is referred to as retrofitting. Retrofitting of a structure is not the same as repair or rehabilitation. Repair refers to partial improvement of the degraded strength of a structure after an earthquake, in fact, it is only a cosmetic treatment. Rehabilitation is a treatment of the structure aimed at achieveing the original strength of the structure after it has deteriorated and suffered damage. Retrofitting means structural strengthening of a structure to a predefined performance level irrespective of whether the structure is damaged or not.
The repair or strengthening of an existing structure is a greater challenge for a civil engineer compared to designing or constructing a new one. A specfic technology has to be designed and developed to re-establish the strength of damaged structures, and to improve the performance for new functions of old undamaged structures. Thus, the technique to be used should be simple in implementation; offer better performance when handled by less experienced workers and must use materials that are readily available, durable, strong and economical. Retrofitting of individual members is referred to as ‘local retrofitting’. For this, a large number of techniques are being used including replacement technique, removal, injection technique, shotcreting and plate bonding etc. Amongst all the above techniques the plate bonding technique is found to be the most efficient and suitable method for retrofitting purposes. In the plate bonding technique, Ferrocement Plates, Fiber Reinforced Polymer (FRP) Plates, Polymer modified concrete and mortar (PMC/PMM) and Steel plates are most commonly used for retrofitting. Of this techniques the use of Fiber Reinforced Polymer (FRP) Plates has gained significant popularity in the last two to three decades. But this technique is costly and requires skilled labour. Various authors have suggested the use of ferrocement jacketing as a more attractive choice in place of FRP plate bonding technique due to its easy application, lesser weight, higher impermeability, improved tensile strength, economical use, and long life term performance.
In the present study, an effort has been made to study the effect of ferrocement jacketing on the strength of retrofitted beam-column joints. The studies have been carried out for various parameters like number of wire mesh layers and their orientation in the ferrocement jackets and initial stress levels of the beam-column joints. The effect of these parameters on the strength of reinforced concrete beam-column joints initially stressed to pre-determined levels, and subsequently retrofitted with ferrocement jackets was investigated. A similar set of beam-column joints was also retrofitted using two layers of CFRP jackets with an orientation of 45° to the longitudinal axis of the joint, to study the behavior of such beam-column joints.
Subsequently, a 3D nonlinear finite element (FE) model using software ATENA-3D was used to validate the experimental results. Comparison between the finite element and experimental results confirms a reasonable accuracy of the proposed model.
The test results showed that retrofitting beam-column joints with different layers of wire mesh in the ferrocement jackets and two layers of CFRP jacketing significantly increased the ultimate and yield load carrying capacity, stiffness of all the joints stressed to various levels, establishing the efficacy of using the material for retrofitting. The use of ferrocement and CFRP jacketing for retrofitting of initially stressed beam-column joints helped to regain full strength even if stressed to 85% of the ultimate load. Due to the strengthening of beam-column joints of control specimens, the failure of the retrofitted beam-column joint specimens shifted from the joint region to the beam ends in the retrofitted specimens. This would help in preventing progressive collapse of the structure. The retrofitting of the beam-column joints may thus shift the failure from the joint to the beam end to obtain a weak beam- strong joint failure pattern. The comparison between the load-deflection results obtained from ATENA 3D and the experimental study shows that the ATENA 3D results agree reasonably with the experimental results. The variation of experimental and FEM (ATENA 3D) load results for the control as well as retrofitted specimens was within ± 10%. The element modeling (ferrocement and CFRP) showed higher values as compared to the experimentally obtained values.Department of Civil Engineering, Thapar University, Patial
Analysis of Power Line Communication Channel Model using Communication Techniques
No full textWith the advent of technology, human dependency on power (electricity) and communication has grown beyond leaps and bounds. Many efforts have been made to continuously improve and increase the efficiency in both areas. Power Line Communication (PLC) is a technology where power lines or transmission lines are being used for communication purposes along with transmitting electrical energy. Because the power grid is already in place, the PLC has the obvious advantage of reducing communication infrastructure cost. However, the power grid is designed optimally for power delivery (not the data). The power transmission line generally appears as a harsh environment for the low-power high-frequency communication signals. In order to evaluate the performance of PLC, this paper simulates a practical multipath power line communication channel model and provides the Bit Error Rate (BER) vs signal-to-noise ratio (SNR) curves for orthogonal frequency division multiplexing
Semiotic Engineering (I-System) Thumb Rule to Fill the Gap between Formal Principles & Practical Realizations of Textile, Garment & Fashion Technology
No full textWord Sense Disambiguation for Punjabi Language
No full textThe recent eruption of data in different natural languages on the internet has necessitated the
development of Natural Language Processing (NLP) tasks. The major impediments in the
development and implementation of NLP are the scarcity of the standard datasets, knowledge
resources, language tools and ambiguity resolution. Word Sense Disambiguation (WSD) is a
critical and essential task for machine translation, information retrieval, question answering
and sentiment analysis, etc. NLP tasks.
The objective of WSD is to automatically select the appropriate sense of an ambiguous word
based on the context of the word. The WSD process identifies the different senses for every
word relevant to the text or discourse under consideration from the sense inventories such as
dictionaries, thesaurus, ontologies, and WordNet. Then it involves a mean to assign the
appropriate sense to each occurrence of a word in context. Thus WSD needs the representation
of common sense and encyclopaedic knowledge to resolve the sense of ambiguous words.
Recognizing the proper sense of a word in context by a computer is defined as an AI-complete
complexity problem. There are two different types of WSD, namely targeted WSD and allwords WSD. The targeted WSD resolves the ambiguity of an ambiguous target word, usually
occurring one per sentence. The all-words WSD disambiguates all open-class words (noun,
adverb, verb, adjectives) in a text. In this research work, targeted WSD was implemented.
India is a multilingual country having 22 national languages. Interlanguage processing tasks
like machine translation, question answering, sentiment analysis, cross-lingual search, etc., are
highly applicable problems in India. The Punjabi language is the official language of the Indian
state of Punjab, and it is the world’s 10th most widely spoken language. The Punjabi diasporas
are present globally, and there is a need for the Punjabi NLP tasks to connect them to the
Punjabi language successfully. It motivated us to explore the field of WSD for the Punjabi
language. WSD has been successfully designed and developed for the English language.
However, there are many differences in the language structure of English and Punjabi, which
arise different challenges while performing sense disambiguation on the Punjabi text dataset.
In this research work, the systematic review has explicitly portrayed WSD in Indian languages.
A review methodology has followed with the help of the framed research questions. The
renowned electronic databases and the topmost conferences were explored to include the
relevant studies of WSD for Indian languages. The existing status of the WSD for Indianiv
languages has categorised as per the different families of the Indian languages. The evolution
of WSD for Indian languages and their publication time is reported. This research has reviewed
and analysed the WSD for Indian languages based on the techniques, knowledge resources and
evaluation methods. The review of the standard Senseval/Semeval evaluation workshops for
WSD field has been presented. The findings of this research work, such as the available raw
corpora, sense tagged corpus, dictionaries, WordNet and pre-processing linguistic tools of
different Indian languages, will help the researchers. The comprehensive survey presented in
this research work would assist the researchers in choosing the most suitable WSD in the
specific domain and the pertinent future directions. The availability of the Punjabi WordNet
has motivated us to implement the Punjabi WSD
Soft Computing Techniques in a Distributed Environment
No full textThis present investigation deals with the ability of soft computing techniques, in particular, artificial neural networks in solving vibration and system identification problems. Artificial neural networks are also called parallel distributed systems because they are composed of a series of interconnected processing elements that operate in parallel. New training algorithms have been developed to train artificial neural networks and those are applied to the vibration analysis of structural members as well as system identification problems of structural dynamics for partially known and completely unknown systems. The structural problem is continually acquiring greater importance in modern science and technology and being solved with the help of different analytical and approximate methods. Structural members are often encountered in several engineering applications and their use in machine design, aeronautical engineering, nuclear reactor technology, naval structures, and earthquake resistance structures are very common. The system identification problem is an area of importance in structural engineering and is used to improve dynamic modeling capabilities for civil infrastructure systems such as high-rise building, bridges and dams.
The main contributions of this study are the following.
• New artificial neural network learning algorithms developed which make use of the coefficients of linear and nonlinear regression polynomials, including single and multiple variables, as training weights. These polynomials depend upon the ANN architecture to be considered for a particular problem. The proposed algorithm is henceforth abbreviated as RBNN (Regression Based Neural
Network). In these models, the number of neurons in the hidden layer may be fixed depending upon the required polynomial degree.
• This RBNN model is applied to estimate the vibration characteristics for free
vibration of elastic plates with different boundary conditions at the edges.
• The proposed algorithm is also applied to partially known and partially unknown system identification problems. Both multi-input single-output and multi-input multi-output cases have been considered for the analysis and simulation.
The whole range of the subject in this thesis is covered in six chapters, which deal with brief discussion of the existing soft computing techniques, vibration and system identification studies and the proposed new algorithms. Then the thesis investigates the xiv application of these techniques to continuous and discrete systems in terms of vibration and inverse vibration analysis. The reliability, efficiency and powerfulness of the
proposed techniques are also discussed by comparing the present with known results in special cases
Removal of VOCs from waste gas streams by a hollow fiber permeator
No full textRemoval of various VOCs from air/nitrogen feed streams using a novel hollow fiber membrane was studied. Hollow Fiber Module (HFM) used had composite silicone membranes wherein an ultrathin (~ 1μm), nonporous silicone rubber membrane layer had been plasma polymerized on a porous (porosity: 0.4) polypropylene substrate. VOCs studied were toluene, methanol, acetone, methylene chloride and hexane. Primary focus was on single VOCs, although separation of VOC mixtures was also briefly studied. HFM was found to be extremely effective in removing various VOCs from feed streams. Removal of 90-99 % of various VOCs was achieved at low feed flow rates and high inlet VOC concentrations. The membrane exhibited high selectivities for VOC over nitrogen/air. The VOC permeance was found to be dependent on the VOC concentration. Tube-side feed and shell-side feed modes of operation were analyzed for methanol and toluene; it was observed that tube-side feed mode gives better VOC separations. A mathematical model was developed and numerically simulated to explain the observed VOC (toluene and methanol) separation behavior of HFM. The model was able to explain the experimental results reasonably well. Removal of VOC (acetone) from a high pressure gas was also studied. HFM was also successful in separating a mixture of VOCs (toluene, methanol, acetone) from a nitrogen feed stream
BER Performance of Alamouti with VBLAST Detection Schemes over MIMO System
No full textMultiple-Input Multiple-Output (MIMO) systems as a means to combat fading in wireless channels. MIMO allows higher throughput, diversity gain and interference reduction. In this paper, we analyze the Bit Error Rate (BER) performance of the Alamouti Space Time Block Code with V-BLAST (Vertical Bell Laboratories Layered Space-Time) over MIMO system. Basic idea in this scheme is to improve the BER performance of systems. V-BLAST algorithm offers highly better error performance than conventional linear receivers and still has low complexity. The simulated results are based on different modulations, such as BPSK, 4-QAM and 16-QAM over Rayleigh fading channels.
DOI: 10.17762/ijritcc2321-8169.150511
- …
