Text recognition approaches for indoor robotics : a comparison

This paper evaluates the performance of different text recognition techniques for a mobile robot in an indoor (university campus) environment. We compared four different methods: our own approach using existing text detection methods (Minimally Stable Extremal Regions detector and Stroke Width Transform) combined with a convolutional neural network, two modes of the open source program Tesseract, and the experimental mobile app Google Goggles. The results show that a convolutional neural network combined with the Stroke Width Transform gives the best performance in correctly matched text on images with single characters whereas Google Goggles gives the best performance on images with multiple words. The dataset used for this work is released as well.

Text mining with semantic annotation : using enriched text representation for entity-oriented retrieval, semantic relation identification and text clustering

This project is a step forward in the study of text mining where enhanced text representation with semantic information plays a significant role. It develops effective methods of entity-oriented retrieval, semantic relation identification and text clustering utilizing semantically annotated data. These methods are based on enriched text representation generated by introducing semantic information extracted from Wikipedia into the input text data. The proposed methods are evaluated against several start-of-art benchmarking methods on real-life data-sets. In particular, this thesis improves the performance of entity-oriented retrieval, identifies different lexical forms for an entity relation and handles clustering documents with multiple feature spaces.

Can skin cancer prevention and early detection be improved via mobile phone text messaging? a randomised, attention-control trial.

Objective. To test the impact of a theory-based, SMS (text message)-delivered behavioural intervention (Healthy Text) targeting sun protection or skin self-examination behaviours compared to attention-control. Method. Overall, 546 participants aged 18–42 years were randomised using a computer-generated number list to the skin self-examination (N = 176), sun protection (N = 187), or attention-control (N = 183) text messages group. Each group received 21 text messages about their assigned topic over 12 months (12 weekly messages for three months, then monthly messages for the next nine months). Data was collected via telephone survey at baseline, three-, and 12-months across Queensland from January 2012 to August 2013. Results. One year after baseline, the sun protection (mean change 0.12; P = 0.030) and skin self-examination groups (mean change 0.12; P = 0.035) had significantly greater improvement in their sun protection habits (SPH) index compared to the attention-control group (reference mean change 0.02). The increase in the proportion of participants who reported any skin self-examination from baseline to 12 months was significantly greater in the skin self-examination intervention group (103/163; 63%; P < 0.001) than the sun protection (83/173; 48%), or attention-control (65/165; 36%) groups. There was no significant effect of the intervention for participants who self-reported whole-body skin self-examination, sun tanning behaviour, or sunburn behaviours. Conclusion. The Healthy Text intervention was effective in inducing significant improvements in sun protection and any type of skin self-examination behaviours.

Using automated text analysis to evaluate students’ conceptual understanding

Background: A major challenge for assessing students’ conceptual understanding of STEM subjects is the capacity of assessment tools to reliably and robustly evaluate student thinking and reasoning. Multiple-choice tests are typically used to assess student learning and are designed to include distractors that can indicate students’ incomplete understanding of a topic or concept based on which distractor the student selects. However, these tests fail to provide the critical information uncovering the how and why of students’ reasoning for their multiple-choice selections. Open-ended or structured response questions are one method for capturing higher level thinking, but are often costly in terms of time and attention to properly assess student responses. Purpose: The goal of this study is to evaluate methods for automatically assessing open-ended responses, e.g. students’ written explanations and reasoning for multiple-choice selections. Design/Method: We incorporated an open response component for an online signals and systems multiple-choice test to capture written explanations of students’ selections. The effectiveness of an automated approach for identifying and assessing student conceptual understanding was evaluated by comparing results of lexical analysis software packages (Leximancer and NVivo) to expert human analysis of student responses. In order to understand and delineate the process for effectively analysing text provided by students, the researchers evaluated strengths and weakness for both the human and automated approaches. Results: Human and automated analyses revealed both correct and incorrect associations for certain conceptual areas. For some questions, that were not anticipated or included in the distractor selections, showing how multiple-choice questions alone fail to capture the comprehensive picture of student understanding. The comparison of textual analysis methods revealed the capability of automated lexical analysis software to assist in the identification of concepts and their relationships for large textual data sets. We also identified several challenges to using automated analysis as well as the manual and computer-assisted analysis. Conclusions: This study highlighted the usefulness incorporating and analysing students’ reasoning or explanations in understanding how students think about certain conceptual ideas. The ultimate value of automating the evaluation of written explanations is that it can be applied more frequently and at various stages of instruction to formatively evaluate conceptual understanding and engage students in reflective

User preferences for text message-delivered skin cancer prevention and early detection

Evidence is needed for the acceptability and user preferences of receiving skin cancer-related text messages. We prepared 27 questions to evaluate attitudes, satisfaction with program characteristics such as timing and spacing, and overall satisfaction with the Healthy Text program in young adults. Within this randomised controlled trial (age 18-42 years), 546 participants were assigned to one of three Healthy Text message groups; sun protection, skin self-examination, or attention-control. Over a 12-month period, 21 behaviour-specific text messages were sent to each group. Participants’ preferences were compared between the two interventions and control group at the 12-month follow-up telephone interview. In all three groups, participants reported the messages were easy to understand (98%), provided good suggestions or ideas (88%), and were encouraging (86%) and informative (85%) with little difference between the groups. The timing of the texts was received positively (92%); however, some suggestions for frequency or time of day the messages were received from 8% of participants. Participants in the two intervention groups found their messages more informative, and triggering behaviour change compared to control. Text messages about skin cancer prevention and early detection are novel and acceptable to induce behaviour change in young adults.

Relevance feature discovery for text mining

It is a big challenge to guarantee the quality of discovered relevance features in text documents for describing user preferences because of large scale terms and data patterns. Most existing popular text mining and classification methods have adopted term-based approaches. However, they have all suffered from the problems of polysemy and synonymy. Over the years, there has been often held the hypothesis that pattern-based methods should perform better than term-based ones in describing user preferences; yet, how to effectively use large scale patterns remains a hard problem in text mining. To make a breakthrough in this challenging issue, this paper presents an innovative model for relevance feature discovery. It discovers both positive and negative patterns in text documents as higher level features and deploys them over low-level features (terms). It also classifies terms into categories and updates term weights based on their specificity and their distributions in patterns. Substantial experiments using this model on RCV1, TREC topics and Reuters-21578 show that the proposed model significantly outperforms both the state-of-the-art term-based methods and the pattern based methods.

Analysing reflective text for learning analytics : an approach using anomaly recontextualisation

Reflective writing is an important learning task to help foster reflective practice, but even when assessed it is rarely analysed or critically reviewed due to its subjective and affective nature. We propose a process for capturing subjective and affective analytics based on the identification and recontextualisation of anomalous features within reflective text. We evaluate 2 human supervised trials of the process, and so demonstrate the potential for an automated Anomaly Recontextualisation process for Learning Analytics.

Text mining for screening efficiency? Testing within a Cochrane public health review

Background The requirement for dual screening of titles and abstracts to select papers to examine in full text can create a huge workload, not least when the topic is complex and a broad search strategy is required, resulting in a large number of results. An automated system to reduce this burden, while still assuring high accuracy, has the potential to provide huge efficiency savings within the review process. Objectives To undertake a direct comparison of manual screening with a semi‐automated process (priority screening) using a machine classifier. The research is being carried out as part of the current update of a population‐level public health review. Methods Authors have hand selected studies for the review update, in duplicate, using the standard Cochrane Handbook methodology. A retrospective analysis, simulating a quasi‐‘active learning’ process (whereby a classifier is repeatedly trained based on ‘manually’ labelled data) will be completed, using different starting parameters. Tests will be carried out to see how far different training sets, and the size of the training set, affect the classification performance; i.e. what percentage of papers would need to be manually screened to locate 100% of those papers included as a result of the traditional manual method. Results From a search retrieval set of 9555 papers, authors excluded 9494 papers at title/abstract and 52 at full text, leaving 9 papers for inclusion in the review update. The ability of the machine classifier to reduce the percentage of papers that need to be manually screened to identify all the included studies, under different training conditions, will be reported. Conclusions The findings of this study will be presented along with an estimate of any efficiency gains for the author team if the screening process can be semi‐automated using text mining methodology, along with a discussion of the implications for text mining in screening papers within complex health reviews.

Injury narrative text classification using factorization model

Narrative text is a useful way of identifying injury circumstances from the routine emergency department data collections. Automatically classifying narratives based on machine learning techniques is a promising technique, which can consequently reduce the tedious manual classification process. Existing works focus on using Naive Bayes which does not always offer the best performance. This paper proposes the Matrix Factorization approaches along with a learning enhancement process for this task. The results are compared with the performance of various other classification approaches. The impact on the classification results from the parameters setting during the classification of a medical text dataset is discussed. With the selection of right dimension k, Non Negative Matrix Factorization-model method achieves 10 CV accuracy of 0.93.

Decision boundary setting and classifier combination for text classification

This thesis presents a promising boundary setting method for solving challenging issues in text classification to produce an effective text classifier. A classifier must identify boundary between classes optimally. However, after the features are selected, the boundary is still unclear with regard to mixed positive and negative documents. A classifier combination method to boost effectiveness of the classification model is also presented. The experiments carried out in the study demonstrate that the proposed classifier is promising.

Deriving word association networks from text corpora

This article presents and evaluates a model to automatically derive word association networks from text corpora. Two aspects were evaluated: To what degree can corpus-based word association networks (CANs) approximate human word association networks with respect to (1) their ability to quantitatively predict word associations and (2) their structural network characteristics. Word association networks are the basis of the human mental lexicon. However, extracting such networks from human subjects is laborious, time consuming and thus necessarily limited in relation to the breadth of human vocabulary. Automatic derivation of word associations from text corpora would address these limitations. In both evaluations corpus-based processing provided vector representations for words. These representations were then employed to derive CANs using two measures: (1) the well known cosine metric, which is a symmetric measure, and (2) a new asymmetric measure computed from orthogonal vector projections. For both evaluations, the full set of 4068 free association networks (FANs) from the University of South Florida word association norms were used as baseline human data. Two corpus based models were benchmarked for comparison: a latent topic model and latent semantic analysis (LSA). We observed that CANs constructed using the asymmetric measure were slightly less effective than the topic model in quantitatively predicting free associates, and slightly better than LSA. The structural networks analysis revealed that CANs do approximate the FANs to an encouraging degree.

Centroid training to achieve effective text classification

Traditional text classification technology based on machine learning and data mining techniques has made a big progress. However, it is still a big problem on how to draw an exact decision boundary between relevant and irrelevant objects in binary classification due to much uncertainty produced in the process of the traditional algorithms. The proposed model CTTC (Centroid Training for Text Classification) aims to build an uncertainty boundary to absorb as many indeterminate objects as possible so as to elevate the certainty of the relevant and irrelevant groups through the centroid clustering and training process. The clustering starts from the two training subsets labelled as relevant or irrelevant respectively to create two principal centroid vectors by which all the training samples are further separated into three groups: POS, NEG and BND, with all the indeterminate objects absorbed into the uncertain decision boundary BND. Two pairs of centroid vectors are proposed to be trained and optimized through the subsequent iterative multi-learning process, all of which are proposed to collaboratively help predict the polarities of the incoming objects thereafter. For the assessment of the proposed model, F1 and Accuracy have been chosen as the key evaluation measures. We stress the F1 measure because it can display the overall performance improvement of the final classifier better than Accuracy. A large number of experiments have been completed using the proposed model on the Reuters Corpus Volume 1 (RCV1) which is important standard dataset in the field. The experiment results show that the proposed model has significantly improved the binary text classification performance in both F1 and Accuracy compared with three other influential baseline models.

[await full-text]Evaluation of QAS Workplace Health and Safety Action Plan

This report identifies the outcomes of a program evaluation of the five year Workplace Health and Safety Strategy (2012-2017), specifically, the engagement component within the Queensland Ambulance Service. As part of the former Department of Community Safety, their objective was to work towards harmonising the occupational health and safety policies and process to improve the workplace culture. The report examines and assess the process paths and resource inputs into the strategy, provides feedback on progress to achieving identified goals as well as identify opportunities for improvements and barriers to progress. Consultations were held with key stakeholders within QAS and focus groups were facilitated with managers and health and safety representatives of each Local Area Service Network.