Mining condensed sets of frequent episodes with more accurate frequencies from complex sequences

Many previous approaches to frequent episode discovery only accept simple sequences. Although a recent approach has been able to nd frequent episodes from complex sequences, the discovered sets are neither condensed nor accurate. This paper investigates the discovery of condensed sets of frequent episodes from complex sequences. We adopt a novel anti-monotonic frequency measure based on non-redundant occurrences, and dene a condensed set, nDaCF (the set of non-derivable approximately closed frequent episodes) within a given maximal error bound of support. We then introduce a series of effective pruning strategies, and develop a method, nDaCF-Miner, for discovering nDaCF sets. Experimental results show that, when the error bound is somewhat high, the discovered nDaCF sets are two orders of magnitude smaller than complete sets, and nDaCF-miner is more efficient than previous mining approaches. In addition, the nDaCF sets are more accurate than the sets found by previous approaches.

The letter contrast sensitivity test : clinical evaluation of a new design

PURPOSE. To compare the reliability, validity, and responsiveness of the Mars Letter Contrast Sensitivity (CS) Test to the Pelli-Robson CS Chart.

METHODS. One eye of 47 normal control subjects, 27 patients with open-angle glaucoma, and 17 with age-related macular degeneration (AMD) was tested twice with the Mars test and twice with the Pelli-Robson test, in random order on separate days. In addition, 17 patients undergoing cataract surgery were tested, once before and once after surgery.

RESULTS. The mean Mars CS was 1.62 log CS (0.06 SD) for normal subjects aged 22 to 77 years, with significantly lower values in patients with glaucoma or AMD (P < 0.001). Mars test-retest 95% limits of agreement (LOA) were ±0.13, ±0.19, and ±0.24 log CS for normal, glaucoma, and AMD, respectively. In comparison, Pelli-Robson test-retest 95% LOA were ±0.18, ±0.19, and ±0.33 log CS. The Spearman correlation between the Mars and Pelli-Robson tests was 0.83 (P < 0.001). However, systematic differences were observed, particularly at the upper-normal end of the range, where Mars CS was lower than Pelli-Robson CS. After cataract surgery, Mars and Pelli-Robson effect size statistics were 0.92 and 0.88, respectively.

CONCLUSIONS. The results indicate the Mars test has test-retest reliability equal to or better than the Pelli-Robson test and comparable responsiveness. The strong correlation between the tests provides evidence the Mars test is valid. However, systematic differences indicate normative values are likely to be different for each test. The Mars Letter CS Test is a useful and practical alternative to the Pelli-Robson CS Chart.

Reliability and validity of the Melbourne edge test and high/low contrast visual acuity chart

Purpose. The purposes of the study were to investigate the test-retest reliability and the validity of new versions of the Melbourne Edge Test (MET) and the High/Low Contrast Visual Acuity (H/LCVA) chart and to investigate the agreement between the original and new versions.

Methods. The MET original photographic version, MET new light box version, H/LCVA Chart original photographic version, H/LCVA Chart new printed version, and the Pelli-Robson chart were administered twice to one eye of 22 subjects with low vision and 20 soft contact lenses wearers.

Results. For the low vision group, the test-retest 95% limits of agreement were ±5.2 dB for the MET new light box version and ±0.39 logarithm of the minimum angle of resolution (logMAR) for the LCVA component of the H/LCVA new printed version. For the soft contact lens group, the test-retest 95% limits of agreement were ±2.1 dB for the MET new light box version and ± 0.26 logMAR for the LCVA component of the H/LCVA new printed version. Moderate to high correlations were obtained between contrast sensitivity tests, thus providing evidence of validity. Scores obtained for the new test versions were significantly higher than the original versions (p < 0.01).

Conclusions. Of all the tests administered, the MET original photographic version and the Pelli-Robson Chart had the highest test-retest reliability for the low vision group. For the soft contact lens group, the H/LCVA original version (low contrast letters, 18% Weber) and the Pelli-Robson Chart had the highest reliability.

A comparative analysis of decision trees vis-à-vis other computational data mining techniques in automotive insurance fraud detection

The development and application of computational data mining techniques in financial fraud detection and business failure prediction has become a popular cross-disciplinary research area in recent times involving financial economists, forensic accountants and computational modellers. Some of the computational techniques popularly used in the context of - financial fraud detection and business failure prediction can also be effectively applied in the detection of fraudulent insurance claims and therefore, can be of immense practical value to the insurance industry. We provide a comparative analysis of prediction performance of a battery of data mining techniques using real-life automotive insurance fraud data. While the data we have used in our paper is US-based, the computational techniques we have tested can be adapted and generally applied to detect similar insurance frauds in other countries as well where an organized automotive insurance industry exists.

Advanced pattern mining for complex sata analysis

The thesis has researched a set of critical problems in data mining and has proposed four advanced pattern mining algorithm to discover the most interesting and useful data patterns highly relevant to the user’s application targets from the data is represented in complex structures.

EFP-M2 : efficient model for mining frequent patterns in transactional database

Discovering frequent patterns plays an essential role in many data mining applications. The aim of frequent patterns is to obtain the information about the most common patterns that appeared together. However, designing an efficient model to mine these patterns is still demanding due to the capacity of current database size. Therefore, we propose an Efficient Frequent Pattern Mining Model (EFP-M2) to mine the frequent patterns in timely manner. The result shows that the algorithm in EFP-M2l is outperformed at least at 2 orders of magnitudes against the benchmarked FP-Growth.