Weighting Methods for Rare Event Identification From Imbalanced Datasets.

In machine learning, we often face the situation where the event we are interested in has very few data points buried in a massive amount of data. This is typical in network monitoring, where data are streamed from sensing or measuring units continuously but most data are not for events. With imbalanced datasets, the classifiers tend to be biased in favor of the main class. Rare event detection has received much attention in machine learning, and yet it is still a challenging problem. In this paper, we propose a remedy for the standing problem. Weighting and sampling are two fundamental approaches to address the problem. We focus on the weighting method in this paper. We first propose a boosting-style algorithm to compute class weights, which is proved to have excellent theoretical property. Then we propose an adaptive algorithm, which is suitable for real-time applications. The adaptive nature of the two algorithms allows a controlled tradeoff between true positive rate and false positive rate and avoids excessive weight on the rare class, which leads to poor performance on the main class. Experiments on power grid data and some public datasets show that the proposed algorithms outperform the existing weighting and boosting methods, and that their superiority is more noticeable with noisy data.

Online couples mindfulness-based intervention for young breast cancer survivors and their partners: A randomized-control trial.

PURPOSE: Young breast cancer survivors (YBCS) face unique challenges in coping with disease, distress, and relationship concerns. The purposes of this study were to understand the acceptability and feasibility of an online Mindfulness-Based Intervention (MBI) for YBCS and their partners (i.e., Couples Mindfulness-Based Intervention: C-MBI) and to compare the effectiveness of the C-MBI to a closely-matched control, an online MBI for individuals (I-MBI). METHODS: YBCS and their partners were recruited. Couples were randomly assigned to an 8-week C-MBI (couples = 41) or to I-MBI (couples = 36), which included one-hour video modules, a manual, and guided-meditation audios. Both couple members participated in the C-MBI; only the YBCS participated in the control I-MBI. Participants answered surveys about individual- and couple-level functioning at baseline and post-intervention. RESULTS: Online delivery was shown to be feasible and acceptable. For YBCS and their partners, levels of perceived stress, anxiety, depression, and fatigue were lower after the intervention, in both conditions. Unexpectedly, however, participating in the C-MBI appeared to have detrimental effects on dyadic adjustment and relationship quality. CONCLUSION: Although YBCS and their partners reported online delivery was acceptable and benefited well-being, for couple-based MBIs to have benefits for relationship functioning, it may be necessary for couples to have the support of other couples and an instructor. Online delivery may be particularly acceptable and effective for clinical populations, including YBCS. Medical professionals may be more likely to recommend online-MBI programs to cancer survivors, because the programs are of little or no cost.

A mathematical formulation of DNA computation.

DNA computation is to use DNA molecules for information storing and processing. The task is accomplished by encoding and interpreting DNA molecules in suspended solutions before and after the complementary binding reactions. DNA computation is attractive, due to its fast parallel information processing, remarkable energy efficiency, and high storing capacity. Challenges currently faced by DNA computation are: 1) lack of theoretical computational models for applications and 2) high error rate for implementation. This paper attempts to address these problems from mathematical modeling and genetic coding aspects. The first part of this paper presents a mathematical formulation of DNA computation. The model may serve as a theoretical framework for DNA computation. In the second part, a genetic code based DNA computation approach is presented to reduce error rate for implementation, which has been a major concern for DNA computation. The method provides a promising alternative to reduce error rate for DNA computation.