Search

The present study examines the conversational turn-taking patterns in autist-neurotypical talk. The objective of the study is to find out the distinctive features of autist-normal conversations. This study is cross-sectional, descriptive and qualitative in its nature. Recordings are done in anautism center in Lahore for a period of ten days. It is mainly a qualitative study in its nature. Five autistic children of different ages are selected from an autism center in Lahore. The data for the study is collected through video recording of the conversations between autists and speech therapists. The sample is selected through convenient sampling and analysis is done by following the methods of conversation analysis. The results of the analysis highlight certain distinct features of autist child-therapist talk which are not observed in the normal ordinary conversation. However, there is not a total violation of the conversation rules on the part of autists. Moreover, the findings of the research show that conversational patterns in autist-normal conversation are also affected by the chronological age of the autists. Finally, the research concludes that conversation analysis can be used as a tool for the identification of autism.

The focus of this study is the analysis of Punjabi syllablestructure in the light of Generative Phonology Theory. UnderGenerative Phonology, CV phonology theory is used for the analysis of thePunjabi syllable. CV phonology is a three-tiered model explaining thestructure of the syllable. The data for this study is taken from native Punjabispeakers living in Lahore and its surrounding areas. Results of the studysuggest that Punjabi speakers use four main patterns for Syllableconstruction, namely CV, CVC, V and VC.

Objective: To study epidemiology, presentation and management of cases with unilateral proptosis presenting at a tertiary care hospital. Study Design: Case series. Place and Duration of Study: Armed Forces Institute of Ophthalmology, Rawalpindi Pakistan, from Jan 2016 to Dec 2017. Methodology: Retrospective analysis of hospital record of patients with unilateral proptosis was carried out. Data was entered and analyzed in SPSS version 22. Chi square test was used for statistical analysis. Results: Thirty-three cases of unilateral proptosis with a mean age of 41.79 ± 3.87 years were included in the study. Pain was the most common association. Proptosis was mild in 42.4%, moderate in 48.5% and severe in 9.1% cases. In 61% cases, diagnosis was made on excisional or incisional biopsy and on the basis of radio imaging in 39% cases. Most common cause of proptosis found was pleomorphic adenoma (12%). Patients with axial and non-axial proptosis were 48.5% and 51.5% respectively. Fourteen cases (42.5%) were managed surgically and 2 cases (6%) were observed. Total of 22 cases (67%) improved after treatment, 8 cases (24%) went into remission, 2 (6%) deteriorated, and 1 (3%) died in the study period. Conclusion: Unilateral proptosis is a diagnostic challenge. Pertinent history taking and detail examination is the key, but associated signs or symptoms not always point towards correct diagnosis. Tissue biopsy provides a conclusive diagnosis especially where radio imaging is doubtful.

The study examines the impact of Literacy and Numeracy Drive (LND) on the English language learning outcomes and environment at primary sections of rural and urban public schools of district Gujranwala, Punjab, Pakistan. The data were gathered from two public schools: one from a rural area and another from an urban center. The partakers, N=60, served in the controlled and experiment groups. The control group was imparted the prescribed syllabus with the traditional method. In contrast, the experimental group was trained through LND's systematic learning objectives (SLOs), Tablet PCs, specific booklets, and tests. Three pre-tests were conducted before the treatment, and three post-tests were administered after two months of instruction. The study used quasi-experimental methods, and the gathered data were analyzed using SPSS 25. The results show a significant improvement in the outcomes of the experimental group, while no significant change was observed in the case of the control group. The results demonstrate that systematic SOL-based language teaching, technological assistance, and hand training help improve the English language pedagogy. The study suggests that such intervention scan enhance the overall English language teaching environment. As a result,the reach of LND should be extended to other sections of public schools

Automatic incident detection (AID) plays a vital role among all the safety-critical applications under the parasol of Intelligent Transportation Systems (ITSs) to provide timely information to passengers and other stakeholders (hospitals and rescue, police, and insurance departments) in smart cities. Moreover, accurate classification of these incidents with respect to type and severity assists the Traffic Incident Management Systems (TIMSs) and stakeholders in devising better plans for incident site management and avoiding secondary incidents. Most of the AID systems presented in the literature are incident type-specific, i.e., either they are designed for the detection of accident or congestion. While traveling along the road, one may come across different types of traffic incidents, such as accidents, congestion, and reckless driving. This necessitates that the AID system detects and classifies not only all the popular traffic incident types, but severity as well that is associated with these incidents. Therefore, this study aims to propose an efficient incident detection and classification (E-IDC) framework for smart cities, by incorporating the efficacy of model stacking, to classify the incidents with respect to their types and severity levels. The experimental results showed that the proposed E-IDC framework achieved performance gains of 5%–56% in terms of incident severity classification and 1%–14% in terms of incident type classification when applied with different classifiers. We have also applied the Wilcoxon test to benchmark the performance of our proposed framework that reflects the significance of our approach over existing individual incident predictors in terms of severity and type classification. Moreover, it has been observed that the proposed E-IDC framework outperforms the existing ensemble technique, such as XGBoost used for the classification of incidents.

Objective: To study the spectrum of optic neuropathies in patient’s presenting at AFIO Rawalpindi. Study Design: Descriptive study/retrospective study. Place and Duration of Study: Armed Forces Institute of Ophthalmology, Rawalpindi, from Jan 2017 to Dec 2017. Methodology: Retrospective analysis of hospital record of patients coming to neuro-ophthalmic clinic of ArmedForces Institute of Ophthalmology (AFIO), Rawalpindi, from Jan to Dec 2017 was done. Only patients with clearfinal diagnosis of neurophthalmic disease were included. Those with vague diagnosis were excluded. Data wasentered in SPSS version 22 and grouped into variables before analysis. Chi-square test was used for statisticalanalysis. The p

In the present study, hexagonal boron nitride (h-BN) was synthesized from boric acid and melamine by thermal annealing method in a nitrogen atmosphere. The pure h-BN was used as an efficient sorbent for the uptake of Cd2+ ions from the solution phase. The kinetics and sorption studies of metal ions onto the h-BN were carried out in batch adsorption experiments at different temperature, time, pH, sorbent dosage, and concentration of metal ions. The optimum pH for the removal of the Cd2+ ions was found to be pH 7. The effect of temperature showed that the process of Cd2+ sorption remained endothermic in the range of 298 K–328 K. The Lagergren's first and Ho's second kinetic models were tested to interpret the adsorption kinetic data, however the present data was explained well by Ho's model for kinetics. The thermodynamic perameters ΔG, ΔS and ΔH were determined using the available adsorption data at different temperatures. The physicochemical properties of the synthesized product were also characterized before and after adsorption by different analytical techniques like FT-IR, TGA, XRD and Point of Zero Charge (PZC). The morphology of the surface was analyzed with the help of Scanning Electron Microscopy. The h-BN proved to be an efficient adsorbent for the uptake of the Cd2+ ions from aqueous media.

Online learning platforms such as Massive Open Online Course (MOOC), Virtual Learning Environments (VLEs), and Learning Management Systems (LMS) facilitate thousands or even millions of students to learn according to their interests without spatial and temporal constraints. Besides many advantages, online learning platforms face several challenges such as students’ lack of interest, high dropouts, low engagement, students’ self-regulated behavior, and compelling students to take responsibility for settings their own goals. In this study, we propose a predictive model that analyzes the problems faced by at-risk students, subsequently, facilitating instructors for timely intervention to persuade students to increase their study engagements and improve their study performance. The predictive model is trained and tested using various machine learning (ML) and deep learning (DL) algorithms to characterize the learning behavior of students according to their study variables. The performance of various ML algorithms is compared by using accuracy, precision, support, and f-score. The ML algorithm that gives the best result in terms of accuracy, precision, recall, support, and f-score metric is ultimately selected for creating the predictive model at different percentages of course length. The predictive model can help instructors in identifying at-risk students early in the course for timely intervention thus avoiding student dropouts. Our results showed that students’ assessment scores, engagement intensity i.e. clickstream data, and time-dependent variables are important factors in online learning. The experimental results revealed that the predictive model trained using Random Forest (RF) gives the best results with averaged precision =0.60%, 0.79%, 0.84%, 0.88%, 0.90%, 0.92%, averaged recall =0.59%, 0.79%, 0.84%, 0.88%, 0.90%, 0.91%, averaged F-score =0.59%, 0.79%, 0.84%, 0.88%, 0.90%, 0.91%, and average accuracy =0.59%, 0.79%, 0.84%, 0.88%, 0.90%, 0.91% at 0%, 20%, 40%, 60%, 80% and 100% of course length.

Semantic Web is considered as the backbone of web 3.0 and ontologies are an integral part of the Semantic Web. Though an increase of ontologies in different domains is reported due to various benefits which include data heterogeneity, automated information analysis, and reusability, however, finding an appropriate ontology according to user requirement remains cumbersome task due to time and efforts required, context-awareness, and computational complexity. To overcome these issues, an ontology recommendation framework is proposed. The Proposed framework employs text categorization and unsupervised learning techniques. The benefits of the proposed framework are twofold: 1) ontology organization according to the opinion of domain experts and 2) ontology recommendation with respect to user requirement. Moreover, an evaluation model is also proposed to assess the effectiveness of the proposed framework in terms of ontologies organization and recommendation. The main consequences of the proposed framework are 1) ontologies of a corpus can be organized effectively, 2) no effort and time are required to select an appropriate ontology, 3) computational complexity is only limited to the use of unsupervised learning techniques, and 4) due to no requirement of context awareness, the proposed framework can be effective for any corpus or online libraries of ontologies.

The demand of accurate and visually fair images is increasing with the passage of time and bang of the number of digital images especially in the domain of medical and healthcare systems. The visual image quality of modern cameras affected due to edges, textures and sharp structures noise. Though research community has introduced several techniques such as BM3D (Block Matching and 3D) for image denoising. However, edges and texture preservation capabilities remain issues due to hard thresholds values and captured image diversity. In order to address these issues, we propose a new variant of BM3D namely BM3DMA (Block Matching and 3D with Mahalanobis and Adaptive filter) which is employed through the use of Mahalanobis distance measure (for diversity coverage) and adaptive filter (for soft thresholds). We used two widely known datasets consist of set of standard and medical images. We observe 5% to 10% enhancement in the performance of BM3DMA as compared to BM3D in terms of improving the PSNR (Peak Signal to Noise Ratio) value. The promising experimental result indicates the effectiveness of BM3DMA in terms preserving the edge and texture image noise.

The use of deep learning DL techniques for mobile learning is an emerging field aimed at developing methods for finding mobile learners' learning behavior and exploring important learning features. The learning features learning time, learning location, repetition rate, content types, learning performance, learning time duration, and so on act as fuel to DL algorithms based on which DL algorithms can classify mobile learners into different learning groups. In this study, a powerful and efficient m-learning model is proposed based on DL techniques to model the learning process of m-learners. The proposed m-learning model determines the impact of independent learning features on the dependent feature i.e. learners? performance. The m-learning model dynamically and intuitively explores the weights of optimum learning features on learning performance for different learners in their learning environment. Then it split learners into different groups based on features differences, weights, and interrelationships. Because of the high accuracy of the DL technique, it was used to classify learners into five different groups whereas random forest RF ensemble method was used in determining each feature importance in making adaptive m-learning model. Our experimental study also revealed that the m-learning model was successful in helping m-learners in increasing their performance and taking the right decision during the learning flow.

Mobile learning (M-learning) has gained tremendous attention in the educational environment in the past decade. For effective M-learning, it is important to create an efficient M-learning model that can identify the exact requirements of mobile learners (M-learners). M-learning model is composed of features that are generated during M-learners' interaction with mobile devices. For an adaptive M-learning model, not only learning features are required, but it is also important to determine how they differ for various M-learners, their weights, and interrelationship. This study proposes a robust and adaptive M-learning model that is based on machine learning and deep learning (ML/DL) techniques. The proposed M-learning model dynamically explores learning features, their corresponding weights, and association for M-learners. Based on learning features, the M-learning model categorizes M-learners into different performance groups. The M-learning model then provides adaptive content, suggestions, and recommendations to M-learners in order to make learning adaptive and stimulating. For comparative analysis, the prediction accuracy of five baseline ML models was compared with the deep Artificial Neural Network (deep ANN). The results demonstrated that deep ANN and Random Forest (RF) models exhibited better prediction accuracy. Subsequently, both models were selected for developing the M-learning model which included the performance categorization of M-learners under a five-level classification scheme and assigning weights to various features for providing adaptive help and support to M-learners. Our explanatory analysis has shown that behavioral features besides contextual features also influence the learning performance of M-learners. As a direct outcome of this research, more efficient, interactive, and useful mobile learning applications can be developed that accurately predict learning objectives and requirements of diverse M-learners thus helping M-learners in enhancing their study behavior.

It is a knotty task to amicably identify the sporadically changing real-world context information of a learner during M-learning processes. Contextual information varies greatly during the learning process. Contextual information that affects the learner during a learning process includes background knowledge, learning time, learning location, and environmental situation. The computer programming skills of learners improve rapidly if they are encouraged to solve real-world programming problems. It is important to guide learners based on their contextual information in order to maximize their learning performance. In this paper, we proposed a cloud-supported machine learning system (CSMLS), which assists learners in learning practical and applied computer programming based on their contextual information. Learners? contextual information is extracted from their mobile devices and is processed by an unsupervised machine learning algorithm called density-based spatial clustering of applications with noise (DBSCAN) with a rule-based inference engine running on a back-end cloud. CSMLS is able to provide real-time, adaptive, and active learning support to students based on their contextual information characteristics. A total of 150 students evaluated the performance and acceptance of CSMLS for a complete academic semester, i.e. 6 months. Experimental results revealed the threefold success of CSMLS: extraction of students? context information, supporting them in appropriate decision-making, and subsequently increasing their computer programming skills.

Due to the ubiquitous nature of mobile devices, they are now considered as an emerging platform for facilitating both teaching and learning experiences. In this paper, we presented a tailored mobile learning system, namely the Integrated English and Programming Language Learning System (IEPLS), which aims at learning English vocabulary before studying programming language concepts. The IEPLS supports programming language learning in three ways; (a) Recommending to learn specific English vocabulary used in programming language concepts (b) Adaptation to the learning flow of the learner and (c) Motivating and encouraging learners to learn items based on individual learner's performance. The IEPLS was used by one hundred and fifty undergraduate students for six months. Evaluation results revealed the attainment of IEPLS in supporting learners in learning programming languages backed by precise English vocabulary.

Urdu is among the most widely used languages in the world for verbal and written communication. Due to lack of optimized and user friendly native Urdu-script support on various platforms, it is mostly written in Romanized script in soft form. In our research, we have developed a refined Urdu lexicon using tokens with the highest frequency of occurrence in the data set. This data set is basically a raw corpus of colloquial Urdu written in Romanized script. The corpus was collected from volunteer participants who used this language as a mode of communication on the Internet and text massaging. The raw corpus is passed through a series of steps such as Prepossessing, Tokenization and Annotation before passing it to computationally extensive subsequent steps. Edit Distance and K-means Clustering techniques are used for identification of candidate tokens and their potential selection/ inclusion in the refined lexicon. We have also identified most commonly used tokens, candidate tokens and other lingual attributes from the data collected. Based on analysis, we have proposed a computational model for refined colloquial Romanized Urdu lexicon development.

Machine learning (ML) is penetrating in all walks of life and is one of the major driving forces behind the fourth industrial revolution, typically known as Industry 4.0. The purpose of the present study is to review the state-of-the-art ML applications in the biofuels' life cycle stages, i.e., soil, feedstock, production, consumption, and emissions. A keyword search is performed to retrieve relevant articles from the databases of the Web of Science and Google Scholar. ML applications in the soil stage were mostly based on the use of satellite images of land for estimation of biofuels yield or suitability analysis of agricultural land. In the second stage of the life cycle, assessment of rheological properties of the feedstocks and their effect on the quality of biofuels were dominant studies reported in the literature. The production stage included estimation and optimization of quality, quantity, and process conditions. The fuel consumption and emissions stage included analysis of engine performance and estimation of emissions temperature and composition, such as NOx CO, and CO2. This study identified the following trends: dominant ML method, the stage of life cycle getting more usage of ML, the type of data used for the development of the ML-based models, and the stage-wise frequently used input and output variables. The findings of this article are beneficial for academia and industry-related people involved in model development in different stages of biofuel’s life cycle.

The growing public endorsement of social media has changed public life dramatically. Public views and suggestions have now become important for both organizations and individuals. Big data scientists and data mining analysts are increasingly moving their attention toward sentiment analysis because of the growing rate of user-generated contents over microblogging sites. Sentiment analysis is a research field related to computationally identifying public views, feelings, recommendations, opinions and sentiments about focused entities. Research literature shows traces of research work on product and movie reviews for better decision making using big data analysis. Big data analytics offer remarkable opportunities to individuals as well as organizations by providing proficient decision making frameworks and improved forecasting models. The sociopolitical collaboration has gained much attention from online users over the past few years. In this research we analyzed public views, sentiments and opinions shared on social media about a democratic participatory activity called Azadi-March, which was held in Pakistan with participation of online users from all over the world. We carried out computational semantic orientation on public tweets for analyzing public awareness and the effects of online communication through social media over the real world public decision making. We employed unsupervised approach for identification and scoring of tweets. We used lexicon based approach in which annotated lexica are used for scoring verbs, adverbs and other parts of speech. A corpus is used for scoring adjectives and informal opinion indicators. Emoji, exclamatory statements and other additional features are incorporated for supplementary analysis. We noticed that emoticons and NetLingo play significant role in sentiment orientation. Opinion groups are generated from all retrieved tweets and aggregate sentiment weights of opinion groups are computed. The findings of this study indicate that our proposed lexicon based approach outperforms the contemporary machine learning techniques by achieving 86% average accuracy at sentence level sentiment analysis.

As an emerging and promising approach, crowdsourcing-based software development has become popular in many domains due to the participation of talented pool of developers in the contests, and to promote the ability of requesters (or customers) to choose the ‘wining’ solution with respect to their desired quality levels. However, due to lack of a central mechanism for team formation, continuity in the developer’s work on consecutive tasks and risk of noise in submissions of a contest, there is a gap between the requesters of a domain and their quality concerns related to the adaptation of a crowdsourcing-based software development platform. In order to address concerns and aid requesters, we describe three measures; Quality of Registrant Developers (QRD), Quality of Contest (QC) and Quality of Support (QS) to compute and predict the quality of a crowdsourcing-based platform through historical information on its completed tasks. We evaluate the capacity of the QRD, QC and QS as assessors to predict the quality. Subsequently, we implement a crawler to mine the information of completed development tasks from the TopCoder platform to inspect the proposed measures. The promising results of our QRD, QC, and QS measures suggest to use the proposed measures to the requesters and researchers of other domains such as pharmaceutical research and development, in order to investigate and predict the quality of crowdsourcing-based software development platforms.

Progress toward the United Nations 2030 Sustainable Development Goals requires improved information on mortality and causes of death. However, causes of many of the fifty million annual deaths in low- and middle-income countries remain unknown, as most of the deaths occur at home without medical attention. In 2001 India began the Million Death Study in 1.3 million nationally representative households. Nonmedical staff conduct verbal autopsies, which are structured interviews including a half-page narrative in local language of the family's story of the symptoms and events leading to death. Two physicians independently assess each death to arrive at an underlying cause of death. The study has thus far yielded information that substantially altered previous estimates of cause-specific mortality and risk factors in India. Similar robust studies are feasible at low cost in other low- and middle-income countries, particularly if they adopt electronic data management and ensure high quality of fieldwork and physician coding. Nationwide mortality studies enable the quantification of avoidable premature mortality and key risk factors for disease, and provide a practicable method to monitor progress toward the Sustainable Development Goals.

This paper focuses on the word sense disambiguation (WSD) problem in the context of Urdu language. Word sense disambiguation (WSD) is a phenomena for disambiguating the text so that machine (computer) would be capable to deduce correct sense of individual given word(s). WSD is critical for solving natural language engineering (NLE) tasks such as machine translation and speech processing etc. It also increase the performance of other tasks such as text retrieval, document classification and document clustering etc. Research work in WSD has been conducted up to different extents in computationally developed languages of the world. In the context of Urdu language the NLE research in general and the WSD research in particular is still in the infancy stage due to the rich morphological structure of Urdu. In this paper, we use machine learning (ML) approaches such as Bayes net classifier (BN), support vector machine (SVM) and decision tree (DT) for WSD in native script Urdu text. The results shown that BN has better F-measure than SVM and DT. The maximum F-measure of 0.711 over 2.5 million words raw Urdu corpus was recorded for the Bayes net classifier.