Measuring textual semantic similarity has been a subject of intense discussion in NLP and AI for many years. A new area of research has emerged that applies semantic similarity measures within Twitter. However, the development of these measures for the semantic analysis of tweets imposes fundamental challenges. The sparsity, ambiguity, and informality present in social media are hampering the performance of traditional textual similarity measures as "tweets", have special syntactic and semantic characteristics. This paper reviews and evaluates the performance of topological, statistical, and hybrid similarity measures, in the context of Twitter analysis. Furthermore, the performance of each measure is compared against a naïve keyword-based similarity computation method to assess the significance of semantic computation in capturing the meaning in tweets. An experiment is designed and conducted to evaluate the different measures through examining various metrics, including correlation, error rates, and statistical tests on a benchmark dataset. The potential weaknesses of semantic similarity measures in relation to Twitter applications of textual similarity assessment and the research contributions are discussed. This research highlights challenges and potential improvement areas for the semantic similarity of tweets, a resource for researchers and practitioners.

Twitter, a microblogging online social network (OSN), has quickly gained prominence as it provides people with the opportunity to communicate and share posts and topics. Tremendous value lies in automated analysing and reasoning about such data in order to derive meaningful insights, which carries potential opportunities for businesses, users, and consumers. However, the sheer volume, noise, and dynamism of Twitter, imposes challenges that hinder the efficacy of observing clusters with high intra-cluster (i.e. minimum variance) and low inter-cluster similarities. This review focuses on research that has used various clustering algorithms to analyse Twitter data streams and identify hidden patterns in tweets where text is highly unstructured. This paper performs a comparative analysis on approaches of unsupervised learning in order to determine whether empirical findings support the enhancement of decision support and pattern recognition applications. A review of the literature identified 13 studies that implemented different clustering methods. A comparison including clustering methods, algorithms, number of clusters, dataset(s) size, distance measure, clustering features, evaluation methods, and results was conducted. The conclusion reports that the use of unsupervised learning in mining social media data has several weaknesses. Success criteria and future directions for research and practice to the research community are discussed.

Fuzzy sentence semantic similarity measures are designed to be applied to real world problems where a computer system is required to assess the similarity between human natural language and words or prototype sentences stored within a knowledge base. Such measures are often developed for a specific corpus/domain where a limited set of words and sentences are evaluated. As new “fuzzy” measures are developed the research challenge is on how to evaluate them. Traditional approaches have involved rigorous and complex human involvement in compiling benchmark datasets and obtaining human similarity measures. Existing datasets often contain limited fuzzy words and do allow the fuzzy measures to be exhaustively tested. This paper presents an automatic method for the generation of a Multiple Fuzzy Word Dataset (MFWD) from a corpus. A Fuzzy Sentence Pairing Algorithm is used to extract and augment high, medium and low similarity sentence pairs with multiple fuzzy words. Human ratings are collected through crowdsourcing and the MFWD is evaluated using both fuzzy and traditional sentence similarity measures. The results indicated that fuzzy measures returned a higher correlation with human ratings compared with traditional measures.

Short text semantic similarity (STSS) measures are algorithms designed to compare short texts and return a level of similarity between them. However, until recently such measures have ignored perception or fuzzy based words (i.e. very hot, cold less cold) in calculations of both word and sentence similarity. Evaluation of such measures is usually achieved through the use of benchmark data sets comprising of a set of rigorously collected sentence pairs which have been evaluated by human participants. A weakness of these datasets is that the sentences pairs include limited, if any, fuzzy based words that makes them impractical for evaluating fuzzy sentence similarity measures. In this paper, a method is presented for the creation of a new benchmark dataset known as SFWD (Single Fuzzy Word Dataset). After creation the data set is then used in the evaluation of FAST, an ontology based fuzzy algorithm for semantic similarity testing that uses concepts of fuzzy and computing with words to allow for the accurate representation of fuzzy based words. The SFWD is then used to undertake a comparative analysis of other established STSS measures.

A problem in the field of semantic sentence similarity is the inability of sentence similarity measures to accurately represent perception based (fuzzy) words that are commonly used in natural language. This paper presents a new sentence similarity measure that attempts to solve this problem. The new measure, Fuzzy Algorithm for Similarity Testing (FAST) is an ontology based similarity measure that uses concepts of fuzzy and computing with words to allow for the accurate representation of fuzzy based words. Through human experimentation fuzzy sets were created for six categories of words based on their levels of association with particular concepts. These fuzzy sets were then defuzzified and the results used to create new ontological relations between the words. Using these relationships allows for the creation of a new ontology based semantic text similarity algorithm that is able to show the effect of fuzzy words on computing sentence similarity as well as the effect that fuzzy words have on non-fuzzy words within a sentence. Experiments on FAST were conducted using a new fuzzy dataset, the creation of which is described in this paper. The results of the evaluation showed that there was an improved level of correlation between FAST and human test results over two existing sentence similarity measures.

This paper proposes a new model for predicting student learning styles for conversational intelligent tutoring systems (CITS). The learning styles are predicted from behavior cues extracted during conversation obtained during automated CITS tutorials. The heart of the model is a fuzzy rule base determined automatically from existing tutorial data with membership function boundaries optimized by a genetic algorithm. The zero-order Sugeno fuzzy inference model is utilized to predict the Felder and Silverman learning styles in two of the learning style dimensions: perception (sensory-intuitive) and understanding (sequential-global). This work is motivated by the changing nature of both education and learners and the need to provided personalized tutoring on demand. The model is incorporated into an existing CITS and evaluated using undergraduate University students. The experimental results have shown strong predictive accuracy when compared with existing approaches to delivery of personalized tutorials and have received good student feedback.

Student profiling is central to the move from 'one size fits all' computer-aided learning systems to intelligent tutoring systems which adapt to meet the needs of different students. This paper proposes a new method for profiling student learning styles for a conversational intelligent tutoring system (CITS) which utilizes a Mulitlayer Perceptron Artificial Neural Network (MLP-ANN). Throughout an automated conversational tutorial with a CITS, aspects of student behaviour are dynamically captured and input to a Learning Styles Predictor agent to profile an individual's learning style. The proposed method will incorporate a MLP-ANN to combine a set of behaviour traits extracted from the tutoring conversation to improve the accuracy of the learning styles prediction. The paper describes experiments conducted with real students in a live teaching/learning environment for profiling two Felder and Silverman learning styles dimensions. The results show that MLP-ANNs can predict learning styles with an accuracy of 84-89%.

Oscar is a conversational intelligent tutoring system (CITS) which dynamically predicts and adapts to a student's learning style throughout the tutoring conversation. Oscar aims to mimic a human tutor to improve the effectiveness of the learning experience by leading a natural language tutorial and adapting material to suit an individual's learning style. Prediction of learning style is undertaken through capturing independent variables during the conversation. The variable with the highest value determines the individuals learning style. This paper proposes a new method which uses a fuzzy classification tree to build a fuzzy predictive model using these variables which are captured through natural language dialogue Experiments have been undertaken on two of the learning style dimensions: perception (sensory-intuitive) and understanding (sequential-global). Early results show the model has substantially increased the predictive accuracy of the Oscar CITS and discovered some interesting relationships amongst these variables.

Academics who wish to commercialise a novel research concept or prototype `product' face many difficult challenges. These typically include limited business knowledge, the technical verses business language barrier, and the time commitment involved in trying to “sell” the idea to industry and develop one-off prototypes. This paper describes the commercialisation of Goal Orientated Conversational Agents (GO-CA) from initial research concepts through to establishing a UK limited company. GO-CAs have the ability to converse with humans through the use of natural language dialogue in order to achieve a specific task. Such systems are ideal for providing clear and consistent advice 24 hours a day and allow the users to be able to ask questions in natural language which are relevant to their personal situation. The paper briefly describes the novel GO-CA architecture and outlines two developed applications. The pathway to commercialisation which was taken is described and lessons which were learnt are identified.

Intelligent tutoring systems are computer learning systems which personalise their learning content for an individual, based on learner characteristics such as existing knowledge. A recent extension to ITS is to capture student learning styles using a questionnaire and adapt subject content accordingly, however students do not always take the time to complete questionnaires carefully. This paper describes Oscar, a conversational intelligent tutoring system (CITS) which utilises a conversational agent to conduct the tutoring. The CITS aims to mimic a human tutor by dynamically estimating and adapting to a student's learning style during a tutoring conversation. Oscar also offers intelligent solution analysis and problem support for learners. By implicitly modelling the student's learning style during tutoring, Oscar can personalise tutoring to each individual learner to improve the effectiveness of the tutoring. The paper presents the novel methodology and architecture for constructing a CITS. An initial pilot study has been conducted in the domain of tutoring of undergraduate Science and Engineering students using the Index of Learning Styles ILS) model. The experiments to investigate the estimation of learning style have produced encouraging results in the estimation of learning style through a tutoring conversation.

This paper presents the development and commercialisation of a computerised, non-invasive psychological profiling system (named `Silent Talker'), for the analysis of nonverbal behaviour. Nonverbal signals hold rich information about mental, behavioural and/or physical states the detection of which would be beneficial in a wide variety of commercial domains, particularly deception detection and security. Previous attempts to extract individual signals and classify an overall behaviour have been overly complex, time-consuming, costly, biased and error prone. Silent Talker overcomes these problems by the use of Artificial Neural Networks. The testing and validation of the system was undertaken by detecting processes associated with `deception' and `truth' which the system achieved with significant results. Initial academic recognition led to widespread publicity but the commercialisation of the project proved to be far from straightforward. This paper describes the system itself, its potential uses and the commercialisation process to date, concluding with lessons that have been learnt along the way.

Neural networks are a powerful classification technique that are capable of discovering complex relationships between inputs and outputs. This powerful classification ability has meant that neural networks have been widely applied to medical domains. However, neural networks suffer from the so-called blackbox problem: they predict accurately but offer no explanation of how the decision has been derived. This lack of explanation has limited their adoption and has been a concern to the medical community. This paper demonstrates ExTree a rule extraction algorithm being applied to neural networks which have been trained on medical datasets. ExTree extracts from the neural network a decision tree which is a graphical and easily understood representation of a decision process.

The success of fuzzy decision trees when applied to classification problems is usually attributed to the selection and tuning of fuzzy sets to represent the problem domain. The impact of fuzzy inference in combining grades of membership throughout fuzzy trees has not been considered in-depth. A number of parameterized fuzzy operators based on the T-norm model have been proposed but not exploited in practical applications. This paper presents a comparative study which examines a number of T-norm and T-conorms and their application within Fuzzy Decision Trees. The methodology uses a Genetic Algorithm to tune the weights of T-norm operators and optimize fuzzy membership functions simultaneously in fuzzy trees. The paper applies the methodology to two Fuzzy Decision Tree algorithms known as FIA and Fuzzy CHAIRS. Six different T-norm models are investigated across five real world datasets. Experimental results indicate that significant improvements can be made in the performance of fuzzy trees when the most appropriate T-norm is optimised for a specific domain.

Artificial neural networks (ANNs) are a powerful and widely used pattern recognition technique. However, they remain "black boxes" giving no explanation for the decisions they make. This paper presents a new algorithm for extracting a logistic model tree (LMT) from a neural network, which gives a symbolic representation of the knowledge hidden within the ANN. Landwehr's LMTs are based on standard decision trees, but the terminal nodes are replaced with logistic regression functions. This paper reports the results of an empirical evaluation that compares the new decision tree extraction algorithm with Quinlan's C4.5 and ExTree. The evaluation used 12 standard benchmark datasets from the University of California, Irvine machine-learning repository. The results of this evaluation demonstrate that the new algorithm produces decision trees that have higher accuracy and higher fidelity than decision trees created by both C4.5 and ExTree.

Sentence similarity measures play an increasingly important role in text-related research and applications in areas such as text mining, Web page retrieval, and dialogue systems. Existing methods for computing sentence similarity have been adopted from approaches used for long text documents. These methods process sentences in a very high-dimensional space and are consequently inefficient, require human input, and are not adaptable to some application domains. This paper focuses directly on computing the similarity between very short texts of sentence length. It presents an algorithm that takes account of semantic information and word order information implied in the sentences. The semantic similarity of two sentences is calculated using information from a structured lexical database and from corpus statistics. The use of a lexical database enables our method to model human common sense knowledge and the incorporation of corpus statistics allows our method to be adaptable to different domains. The proposed method can be used in a variety of applications that involve text knowledge representation and discovery. Experiments on two sets of selected sentence pairs demonstrate that the proposed method provides a similarity measure that shows a significant correlation to human intuition

We present a hybrid radial basis function (RBF) sigmoid neural network with a three-step training algorithm that utilizes both global search and gradient descent training. The algorithm used is intended to identify global features of an input-output relationship before adding local detail to the approximating function. It aims to achieve efficient function approximation through the separate identification of aspects of a relationship that are expressed universally from those that vary only within particular regions of the input space. We test the effectiveness of our method using five regression tasks; four use synthetic datasets while the last problem uses real-world data on the wave overtopping of seawalls. It is shown that the hybrid architecture is often superior to architectures containing neurons of a single type in several ways: lower mean square errors are often achievable using fewer hidden neurons and with less need for regularization. Our global-local artificial neural network (GL-ANN) is also seen to compare favorably with both perceptron radial basis net and regression tree derived RBFs. A number of issues concerning the training of GL-ANNs are discussed: the use of regularization, the inclusion of a gradient descent optimization step, the choice of RBF spreads, model selection, and the development of appropriate stopping criteria.

Semantic similarity between words is becoming a generic problem for many applications of computational linguistics and artificial intelligence. This paper explores the determination of semantic similarity by a number of information sources, which consist of structural semantic information from a lexical taxonomy and information content from a corpus. To investigate how information sources could be used effectively, a variety of strategies for using various possible information sources are implemented. A new measure is then proposed which combines information sources nonlinearly. Experimental evaluation against a benchmark set of human similarity ratings demonstrates that the proposed measure significantly outperforms traditional similarity measures.

This paper presents a novel approach to combining multiple decision trees, which utilizes the power of fuzzy inference techniques and a back-propagation feed forward neural network (BP-FFNN) to improve the overall classification. Crisp multiple decision trees were originally introduced to improve the performance of single decision trees, which offered a restrictive view of the domain. By combining multiple perspectives of the same domain, the information content is increased and the predictive power of the classifier is improved. A predominant weakness in creating multiple crisp trees is the generation of sharp decision boundaries at every node within all trees. The creation of fuzzy decision forests overcomes this weakness by introducing the concepts of fuzzy theory to soften the decision boundaries. To combine information from all fuzzy trees a unique combination of fuzzy inference is used to generate a series of leaf membership grades, which are then used as input to a BP-FFNN. Comparisons are made between using the fuzzy-neural approach and the use of pure fuzzy inference trees and the results indicate that considerable improvement has been made over crisp multiple decision trees.

The creation of multiple decision trees is a relatively new concept, which aims to improve the predictive power of a single decision tree. The approach is based on the induction of more than one C4.5-type decision tree from the same training sample where each decision tree represents a different view of the same domain resulting in a network of decision tree models. The utilization of multiple decision trees has been shown to lead to an improved performance by combining multiple perspectives of the same domain thus increasing the information content whereas, in comparison, a single decision tree can only represent one restricted view of the domain. One predominant weakness in creating a single tree is the generation of sharp decision boundaries at every node within the tree, which results in small changes in attribute values giving radically different classifications. This problem becomes more apparent with the generation of multiple trees. This paper presents a novel approach of overcoming this weakness through the use of fuzzy decision forests. The approach is based upon the induction of multiple fuzzy decision trees from one training sample, where each tree represents a different view of the data domain. A genetic algorithm (GA) is used to select a series of high performance membership functions, which are then applied to branches within all decision trees in the forest. The GA will in addition optimise a pre-selected fuzzy inference technique, which will assign a degree of strength to the conjunction and disjunction of membership grades within the tree. Considerable improvements in classification accuracy over original single C4.5 (crisp) trees were obtained using two real world data sets.

A network's generalisation is extrapolated from the training examples, i.e. it is defined by the network's model of the training data. Simplifying this model will increase the resistance to noise, inherent in the training data, but will also lose information which may not be noise. This paper discusses the use of Voronoi tessellations for highly detailed data models and the effects of noise. An original algorithm is developed for constructing a competitive layer of k-nearest neighbour type neurons. These neurons are trained to model the decision boundaries and therefore approximate the perfect bisector that is required. The new algorithm is empirically compared with three existing hyperplanic algorithms-neural tree network (NTN), backpropagation (BP), and the entropy net, on three real world continuous data sets.