Lahore Garrison University Research Journal of Computer Science and Information Technology
Not a member yet
227 research outputs found
Sort by
Classifying Urdu Verbs Using Rule Based Approach
No full textTo make dictionaries complete and to keep their size restricted, there is an approach in the linguistic world to equip these dictionaries with morphological information. This module of morphological information is usually known as a morphological analyzer or morphological classifier, which normally contains the complete possible linguistic information about each word for that particular language and it also describes the rules of derivations from the root of a word and its various inflections, respectively. In this work, a classifier for Urdu verbs (CUV) is proposed which is still a challenging research issue, as Urdu is a language of high inflection and derivation. The available stemmers for Urdu do not provide enough information about inflectional and derivational forms of words. Also, morphological classifiers available for Urdu are not worthy of handling various problems and delivering results that prune errors. In our work, a rule based CUV is designed which is able to classify 63 forms of Urdu verbs successfully out of 66. Available Urdu language processing tools are very rare compared to other higher inflectional languages such as German, Turkish, etc., which have competitive morphological classifiers. However, the studies related to Urdu verb morphological classification are identified and a comparative study is presented in this article. In short, this work is a positive contribution to the community, and it provides sufficient information with promising results specifically on inflectional and derivational forms of Urdu verbs
Next-Wave of E-commerce: Mobile Customers Churn Prediction using Machine Learning
No full textWith the swift increase of mobile devices such as personal digital assistants, smartphones and tablets, mobile commerce is broadly considered to be a driving force for the next wave of ecommerce. The power of mobile commerce is primarily due to the anytime-anywhere connectivity and the use of mobile technology, which creates enormous opportunities to attract and engage customers. Many believe that in an era of m-commerce especially in the telecommunication business retaining customers is a big challenge. In the face of an extremely competitive telecommunication industry, the value of acquiring new customers is very much expensive than retaining the existing customer. Therefore, it has become imperative to pay much attention to retaining the existing customers in order to get stabilized in a market comprised of vibrant service providers. In the current market, a number of prevailing statistical techniques for customer churn management are replaced by more machine learning and predictive analysis techniques. In this study, we employed the feature selection technique to identify the most influencing factors in customer churn prediction. We adopt the wrapper-based feature selection approach where Particle Swarm Optimization (PSO) is used for search purposes and different classifiers like Decision Tree (DT), Naïve Bayes, k-NN and Logistic regression is used for evaluation purposes to assess the enactment on optimally sampled and abridged dataset. Lastly, it is witnessed through simulations that our suggested method accomplishes fairly thriving for forecasting churners and hence could be advantageous for exponentially increasing competition in the telecommunication sector
A Survey on Data Security in Cloud Computing Using Blockchain: Challenges, Existing-State-Of-The-Art Methods, And Future Directions
Get PDFCloud computing is one of the ruling storage solutions. However, the cloud computing centralized storage method is not stable. Blockchain, on the other hand, is a decentralized cloud storage system that ensures data security. Cloud environments are vulnerable to several attacks which compromise the basic confidentiality, integrity, availability, and security of the network. This research focus on decentralized, safe data storage, high data availability, and effective use of storage resources. To properly respond to the situation of the blockchain method, we have conducted a comprehensive survey of the most recent and promising blockchain state-of-the-art methods, the P2P network for data dissemination, hash functions for data authentication, and IPFS (InterPlanetary File System) protocol for data integrity. Furthermore, we have discussed a detailed comparison of consensus algorithms of Blockchain concerning security. Also, we have discussed the future of blockchain and cloud computing. The major focus of this study is to secure the data in Cloud computing using blockchain and ease for researchers for further research work
Comparison of Fault Simulation Over Custom Kernel Module Using Various Techniques
Get PDFTo test the behavior of the Linux kernel module, device drivers and file system in a faulty situation, scientists tried to inject faults in different artificial environments. Since the rarity and unpredictability of such events are pretty high, thus the localization and detection of Linux kernel, device drivers, file system modules errors become unfathomable. ‘Artificial introduction of some random faults during normal tests’ is the only known approach to such mystifying problems. A standard method for performing such experiments is to generate synthetic faults and study the effects. Various fault injection frameworks have been analyzed over the Linux kernel to simulate such detection. The following paper highlights the comparison of different approaches and techniques used for such fault injection to test Linux kernel modules that include simulating low resource conditions and detecting memory leaks. The frameworks chosen to be used in these experiments are; Linux Text Project (LTP), KEDR, Linux Fault-Injection (LFI), and SCSI. 
Measuring the Impact of Factors Affecting Game Development in Distributed Software Development
Get PDFA software game is an application that is not only applicable for entertainment purposes but also used in domains like business, education and health care. Software game development is a multidisciplinary process that involves art, sound, artificial intelligence (AI), control systems and human factors which makes it different from traditional software development practice. Distributed software development (DSD) facilitates decentralized zones for the availability of multidisciplinary human resources at less cost. Past studies explored many influencing factors for game development, however, how these factors majorly affect the game development in Distributed Software Development (DSD) environment yet not been studied as per our knowledge. In this research, we not only identified the most influencing factors for game development in DSD but also gauge a relationship matrix between these factors with games’ technical requirements. In our evaluation, we took twenty-nine top-rated animated games to establish a mapping of these factors present in these games. To calculate the variation in a given project budget, we execute Monte-Carlo simulations between the independent variable (influencing factors) and dependent variable (overall cost) that forecast the valuation of each variable impact on the overall nominal cost of the project. Empirical results of our research conclude that among all identified factors, ‘Physical Resources’ and ‘Freelancers’ have a significant impact on the overall project cost. Our research findings quantitatively assist the software project managers to estimate the cost deviations due to influencing factors in Distributed Software Development (DSD) environment.
 
Suicide Rate Predictions In Pakistan By using Neural Networks
Get PDFSuicide is the understudied subject in Pakistan that is a cause of death all over the world. Seventy-fivepercent of suicide occurs in LMIC.In Pakistan information about suicide is limited. The study is about tofind the number of suicide from major cities of Pakistan and then predict the number of suicides by usingNeural Networks Algorithm. About 24639 cases were found in our research from 2001-18 in majorcities of Pakistan. Hanging and poisoning were the most common methods of suicide. The peak age ofsuicide committers was 20-35 included males and females. The lowest number of suicide was inBahawalpur (130 from 2001 to 2018) and the Highest was in Lahore (5925 from 2001 to 2018)
Time Complexity of Color Camera Depth Map Hand Edge Closing Recognition Algorithm
Get PDFThe objective of this paper is to calculate the time complexity of the colored camera depth map hand edge closing algorithm of the hand gesture recognition technique. It has been identified as hand gesture recognition through human-computer interaction using color camera and depth map technique, which is used to find the time complexity of the algorithms using 2D minima methods, brute force, and plane sweep. Human-computer interaction is a very much essential component of most people's daily life. The goal of gesture recognition research is to establish a system that can classify specific human gestures and can make its use to convey information for the device control. These methods have different input types and different classifiers and techniques to identify hand gestures. This paper includes the algorithm of one of the hand gesture recognition “Color camera depth map hand edge recognition” algorithm and its time complexity and simulation on MATLAB
Feature-Based Comparison of Language Transformation Tools
Get PDFCode transformation is the best option while switching from farmer to next technology. Our paper presents a comparative analysis of code transformation tools based on 18 different factors. These factors are Classes, pointers, Access Specifiers, Functions and Exceptions, etc. For this purpose, we have selected varyCode, Telerik, Multi-online converter, and InstantVB. Source Language considered for this purpose is C sharp (C#) and the target language is Visual Basics (VB). Results show that VaryCode is best among the four tools as its converted programs throw fewer errors and require minor changes while running the program
Using codes in place of DNA Sample in Databases to reduce Storage
Get PDFBiological data mainly comprises of Deoxyribonucleic acid (DNA) and protein sequences. These arethe biomolecules that are present in all cells of human beings. Due to the self-replicating property ofDNA, it is a key constituent of genetic material that exists in all breathing creatures. This biomolecule(DNA) comprehends the genetic material obligatory for the operational and expansion of all personifiedlives. To save DNA data of a single person we require 10CD-Rom's. In this paper, A lossless three-phasecompression algorithm is presented for DNA sequences. In the first phase the dataset is segmentedhaving tetra groups and then the resultant genetic sequences are compressed in the form of uniquenumbers (e.g Array Index) and in the second phase binary code is generated on the bases of array indexnumbers and in the last phase the modified version of Run Length Encoding (RLE) is applied on thedataset.The newly proposed technique has been implemented and its performance is also measured on samples.It has achieved the best average compression ratio. After Storing different DNA Samples
Preservation of Privacy of Big Data Using Efficient Anonymization Technique
Get PDFBig data needs to be retained private because of the increase in the amount of data. Data is generated from social networks, organizations and various other ways, which is known as big data. Big data requires large storage as well as high computational power. At every stage, the data needs to be protected. Privacy preservation plays an important role in keeping sensitive information protected and private from any attack. Data anonymization is one of the techniques to anonymize data to keep it private and protected, which includes suppression, generalization, and bucketization. It keeps personal and private data anonymous from being known by others. But when it is implemented on big data, these techniques cause a great loss of information and also fail in defense of the privacy of big data. Moreover, for the scenario of big data, the anonymization should not only focus on hiding but also on other aspects. This paper aims to provide a technique that uses slicing, suppression, and functional encryption together to achieve better privacy of big data with data anonymization