Interactive email filtering: Learning from misclassified examples

Learning from mistakes has proven to be an effective way of learning in the interactive document classifications. In this paper we propose an approach to effectively learning from mistakes in the email filtering process. Our system has employed both SVM and Winnow machine learning algorithms to learn from misclassified email documents and refine the email filtering process accordingly. Our experiments have shown that the training of an email filter becomes much effective and faster

A feasibility study of email communication between the patient's family and the specialist burns team

We investigated whether the parents of burns patients could capture suitable clinical images with a digital camera and add the necessary text information to enable the paediatric burns team to provide follow-up care via email. Four families were involved in the study, each of whom sent regular email consultations for six months. The results were very encouraging. The burns team felt confident that the clinical information in 30 of the 32 email messages (94%) they received was accurate, although in I I of these 30 cases (37%) they stated that there was room for improvement (the quality was nonetheless adequate for clinical decision making). The study also showed that low-resolution images (average size 37 kByte) were satisfactory for diagnosis. Families were able to participate in the service without intensive training and support. The user survey showed that all four families found it easy and convenient to take the digital photographs and to participate in the study. The results suggest that the technique has potential as a low-cost telemedicine service in burns follow-up, and that it requires only modest investment in equipment, training and support.

A review of three years experience using email and videoconferencing for the delivery of post-acute burns care to children in Queensland

A virtual outpatient service has been established in Queensland for the delivery of post-acute burns care to children living in rural and remote areas of the state. The integration of telepaediatrics as a routine service has reduced the need for patient travel to the specialist burns unit situated in Brisbane. We have conducted 293 patient consultations over a period of 3 years. A retrospective review of our experience has shown that post-acute burns care can be delivered using videoconferencing, email and the telephone. Telepaediatric bums services have been valuable in two key areas. The first area involves a programme of routine specialist clinics via videoconference. The second area relates to ad-hoc patient consultations for collaborative management during acute presentations and at times of urgent clinical need. The families of patients have expressed a high degree of satisfaction with the service. Telepaediatric services have helped improve access to specialist services for people living in rural and remote communities throughout Queensland. (C) 2003 Elsevier Ltd and ISBI. All rights reserved.

Recursive filtering of images with symmetric extension

Recursive filters are widely used in image analysis due to their efficiency and simple implementation. However these filters have an initialisation problem which either produces unusable results near the image boundaries or requires costly approximate solutions such as extending the boundary manually. In this paper, we describe a method for the recursive filtering of symmetrically extended images for filters with symmetric denominator. We begin with an analysis of symmetric extensions and their effect on non-recursive filtering operators. Based on the non-recursive case, we derive a formulation of recursive filtering on symmetric domains as a linear but spatially varying implicit operator. We then give an efficient method for decomposing and solving the linear implicit system, along with a proof that this decomposition always exists. This decomposition needs to be performed only once for each dimension of the image. This yields a filtering which is both stable and consistent with the ideal infinite extension. The filter is efficient, requiring less computation than the standard recursive filtering. We give experimental evidence to verify these claims. (c) 2005 Elsevier B.V. All rights reserved.

The use of email in a child and adolescent mental health service: are staff ready?

All staff members of a child and adolescent mental health service were invited to participate in a survey about the use of email. Sixty-two of the 105 staff members responded to the survey, a participation rate of 59%. Of the respondents, 32 were allied health staff, 10 were nurses, seven were administrative staff, six were medical staff, three were operational staff and four were acting in a combination of these roles. The respondents reported extensive work-related email usage and considered that they were confident in using email despite low levels of training. However, they did not feel that they understood the legal and ethical issues involved. Furthermore, there was limited incorporation of email into standard record keeping. The majority of respondents thought that increased use of email would lead to a greater workload, a consequence they considered would probably increase over time. Many commented on the quick and practical use of this medium, but were wary about using email with individuals outside the service organization, especially if it were to contain clinical material. There was low use of email directly with clients, and clinicians were ambivalent about incorporating email into therapy. The results suggest that it is timely to consider the utility and appropriateness of email communication with clients and external service providers, and to formulate guidelines and procedures to ensure the confidentiality of client information and the safety of clients and staff.

Tyre-force estimation by Kalman inverse filtering: applications to off-highway mining trucks

Kalman inverse filtering is used to develop a methodology for real-time estimation of forces acting at the interface between tyre and road on large off-highway mining trucks. The system model formulated is capable of estimating the three components of tyre-force at each wheel of the truck using a practical set of measurements and inputs. Good tracking is obtained by the estimated tyre-forces when compared with those simulated by an ADAMS virtual-truck model. A sensitivity analysis determines the susceptibility of the tyre-force estimates to uncertainties in the truck's parameters.

Email-facilitated qualitative interviews with traumatic brain injury survivors: A new and accessible method

Primary objective: To trial the method of email-facilitated qualitative interviewing with people with traumatic brain injury (TBI). Research design: Qualitative semi-structured email-facilitated interviews. Procedures: Nineteen people (17 severe diagnosis) with a TBI participated in email interviews. Main outcomes and results: Findings indicate that this method facilitates the participation of people with TBI in qualitative interviews. Advantages include increased time for reflection, composing answers and greater control of the interview setting. In addition, the data indicates that people with a TBI are capable of greater insight, reflection and humour than indicated by previous research. Conclusion: Findings indicate that new technologies may advance data collection methods for people with cognitive-linguistic impairments who face participation barriers in face-to-face interviews.

Incremental learning for interactive e-mail filtering

In this article, we propose a framework, namely, Prediction-Learning-Distillation (PLD) for interactive document classification and distilling misclassified documents. Whenever a user points out misclassified documents, the PLD learns from the mistakes and identifies the same mistakes from all other classified documents. The PLD then enforces this learning for future classifications. If the classifier fails to accept relevant documents or reject irrelevant documents on certain categories, then PLD will assign those documents as new positive/negative training instances. The classifier can then strengthen its weakness by learning from these new training instances. Our experiments’ results have demonstrated that the proposed algorithm can learn from user-identified misclassified documents, and then distil the rest successfully.