LANA: An Arabic Conversational Intelligent Tutoring System for Children with Autism Spectrum Disorder Phytochemical and Antimicrobial Studies on Selected Medicinal Plants from the Iraqi Flora Sumayh Aljameel Autism spectrum disorder (ASD) is a lifelong developmental disability, which affects communication skills, social skills and repetitive behaviour. Several studies reported that people with ASD have the ability to deal with technology and computers more effectively than the general population. People with ASD have differences in learning styles. Learning style is a set of factors, attitudes and behaviours that facilitate learning for an individual in a given situation. Researchers have suggested that there are three kinds of learning styles (VAK): through seeing (visual), through hearing (auditory), and through touching an object (kinaesthetic). In Saudi Arabia, the number of children who are diagnosed with autism is increasing. The Saudi Government supports education for these children by inclusion of the children in mainstream schools. However, most of these schools have a lack of specialised or trained teachers to deal with autistic children’s needs. One solution to this problem is to use a virtual tutor to educate children with ASD in mainstream schools. The fundamental research question addressed in this project is can an Arabic Conversational Intelligent Tutoring System (ACITS) adapt to the VAK learning styles for autistic children and enhance their learning. The proposed ACITS architecture uses a combination of Arabic Pattern Matching and Arabic Short Text Similarity to adapt VAK learning style to enhance an individual’s learning in a particular domain (science). The ACITS is aimed at children with autism (10 to 16 years old) who have reached a basic competency with the mechanics of Arabic writing. The new ACITS, known as LANA, will use the learning gain measurement to evaluate the ability of ACITS to personalise the learning of individual through adaptation of learning style in schools. To date, the first prototype has been developed and a preliminary evaluation of the ACITS and its components has been conducted with non-autistic children in order to highlight the strengths and weaknesses. The results show that there is statistically a significant difference between user’s scores with and without using the VAK learning style. In addition, the users who learned using the VAK model are much happier with remembering what they have learned from LANA (91.7%) than users who learned without using the VAK model (50%), indicating that LANA CITS can adapt the user’s learning style and enhance their learning. Moreover, the results show that LANA CITS is effective as an Arabic CA with the majority of conversations leading to the goal of the conversation and the majority of the corrected responses (89%). Current work is in developing prototype 2 of LANA in conjunction with teachers for a pilot with autistic children in Saudi Arabia. Phytochemical and Antimicrobial Studies on Selected Medicinal Plants from the Iraqi Flora Shaymaa Al-Majmaie, Lutfun Nahar, George P. Sharples, Satyajit D. Sarker Herbal medicine is the oldest type of treatment for the maintenance of health and for the prevention of disease, especially in developing countries. Natural products are small molecules, which are secondary metabolites that are not involved in plant growth, development or reproduction. During the last century research was driven mainly towards extraction of plant materials, and isolation of active compounds as potential drug candidates for pharmaceutical industries. In the present study, five plants from the Iraqi flora have been selected because of their traditional medicinal uses, especially in the treatment of infectious diseases. The plants are Ruta chalepensis, Citrus sinensis, Punica granatum, Citrus grandis and Ricinus communis. All of the plant extracts have been tested against Micrococcus luteus, Escherichia coli, Staphylococcus aureus, Pseudomonas aeruginosa and Candida albicans, using agar diffusion assay and the resazurin assay. Analytical studies on the extracts have also been carried out using HPLC. Ruta chalepensis was selected for a detailed investigation of its chemical and antimicrobial properties with the ultimate aim of isolating and purifying the compound responsible for antimicrobial activity.
Risk factors influencing vitamin D status of ethnic minority adults living in the UK Mona Almujaydil With the discovery of skeletal diseases such as osteomalacia and rickets caused by vitamin D deficiency many decades ago, there has been a great interest in the role of vitamin D in human health. In recent years, concerns have been growing about vitamin D deficiency as a significant cause of increased risk of non-skeletal diseases (Holick et al., 2011). Hypovitaminosis D is a serious problem in the UK, due to reduced sunshine exposure and limited dietary sources of vitamin D, coupled with other factors could lead to increased incidence of hypovitaminosis D among the UK population especially amongst ethnic minorities due to their eating habits and high skin pigmentation. Therefore, the purpose of this study is to determine which ethnic minority groups are living in Manchester who are at greater risk of developing vitamin D deficiency. A questionnaire was used to determine diet and lifestyle factors that are associated with a risk of hypovitaminosis D. The estimated mean vitamin D intake by food frequency questionnaire was higher among South Asian with 2.02 µg /day than Arab 1.63µg and Black groups 1.61 µg. while the estimated mean of time outdoors was higher among Asian groups (1.85 hour/day) than other ethnic groups. Other risk factors for vitamin D deficiency included low use of supplements (18%); being overweight or obese (64 % Arab and 39% Black group); smoking and alcohol intake (13.3% Arab, 45.5% Black). The findings will help to determine which groups need more awareness and effective recommendations related to diet and lifestyle that should be adopted or avoided in order to change their lives and attain the adequate level of vitamin D. A Semantic-Based Framework for Social Data Cluster Analysis Noufa Alnajran, Keeley Crockett, David McLean and Annabel Latham This research adds semantic technology to the text mining of Online Social Data (OSN), particularly Twitter microblogging, because simple statistical algorithms that look for keyword occurrences, are proving to be inadequate for the nature of such data. As such, this project sets out to research, develop, and implement a novel approach to clustering Twitter shorttext messages. The effectiveness of this approach relies on an accurate short text semantic similarity (STSS) model that can gauge the degree of semantic equivalence between pairs of short-texts despite the challenges presented in the data. The preliminary evaluation on the SemEval-2014 Twitter dataset revealed weaknesses in current STSS models when applied in the context of Twitter. This is due to the sheer volume, noise, and dynamism of such data that hinder an effective process of analysis and insight extraction. Therefore, the key contributions of the research include the development of a novel semantic-based clustering approach that incorporates a new STSS model for intelligently grouping concepts of microblogging data. Unlike existing studies, this research shall fill the gap of meaning-less keyword-based clustering, and move it towards a semantic-based clustering, which obtains the logic structure and convey meanings. The quantitative methodology of this research is defined through several determined data mining stages. Throughout the first year of research, a thorough literature review on Natural Language Processing (NLP) and text mining techniques have been conducted and a total of 8,234,531 domain-specific (EU Referendum) tweets have been collected via Twitter Streaming API and stored in MongoDB, and a feature extraction function has been implemented to extract a feature set from each tweet. These features include hashtags, mentions, urls, exclamation points or question marks, and named entities (NE), which will form the feature vector along with derived features such as length and part of speech tags (POS). These feature vectors that represent the tweets will contribute in the calculation of the overall (syntactic and semantic) similarity score. Furthermore, various pre-processing methods and NLP techniques have been implemented to reduce the noise and prepare the data for analysis. The processed dataset will be used for the application and evaluation of the developed algorithms. Based on the conducted experiments, hybrid-based STSS will be adapted to develop the new Twitter specific semantic distance measure. A further series of experiments will be conducted to evaluate the new metric and determine the granularity compactness and cohesiveness of the resulting clusters.