Optimal Selection of Reference Set for the NearestNeighbor Classification by Tabu Search

In this paper, a new approach is presented to find the reference set for the nearest neighbor classifier. The optimal reference set, which has minimum sample size and satisfies a certain error rate threshold, is obtained through a Tabu search algorithm. When the error rate threshold is set to zero, the algorithm obtains a near minimal consistent subset of a given training set. While the threshold is set to a small appropriate value, the obtained reference set may compensate the bias of the nearest neighbor estimate. An aspiration criterion for Tabu search is introduced, which aims to prevent the search process from the inefficient wandering between the feasible and infeasible regions in the search space and speed up the convergence. Experimental results based on a number of typical data sets are presented and analyzed to illustrate the benefits of the proposed method. Compared to conventional methods, such as CNN and Dasarathy's algorithm, the size of the reduced reference sets is much smaller, and the nearest neighbor classification performance is better, especially when the error rate thresholds are set to appropriate nonzero values. The experimental results also illustrate that the MCS (minimal consistent set) of Dasarathy's algorithm is not minimal, and its candidate consistent set is not always ensured to reduce monotonically. A counter example is also given to confirm this claim.

STUDY ON A PRECISE ON-SPOT CALIBRATION METHOD FOR DISTRIBUTED STEREO REFERENCE POSITION-SETTING

A high precision method used for on-spot calibration of distributed stereoreference position setting is presented. The high measuring accuracy in stereo reference calibrationis derived from using a high precision water level instrument and an accurate height verniercaliper. It settles the problem of reference calibration effectively and accurately, without usinglarge coordinate measuring machines (CMM). It is more adaptable and precise than traditionalcalibration methods applying theodolites or autocollimators. The error sources of this method areanalyzed in detail and several methods are developed to eliminate the calibration error.Anoptimizing swallowtail-like anchor target is developed. Experiments show that the calibrationaccuracy can be limited within 0.06 mm in the range of 3～5 m and 0.03 mm with optimizing anchortarget. This method can be widely used in on-spot calibration.

A gene family-based method for interspecies comparisons of sequencing-based transcriptomes and its use in environmental adaptation analysis

We describe a new method for sequencing-based cross-species transcriptome comparisons and define a new metric for evaluating gene expression across species using protein-coding families as units of comparison. Using this measure transcriptomes from different species were evaluated by mapping them to gene families and integrating the mapping results with expression data. Statistical tests were applied to the transcriptome evaluation results to identify differentially expressed families. A Perl program named Pro-Diff was compiled to im- plement this method. To evaluate the method and provide an example of its use, two liver EST transcriptomes from two closely related fish that live in different temperature zones were compared. One EST library was from a recent sequencing project of Dissosticus maw- soni, a fish that lives in cold Antarctic sea waters, while the other was newly sequenced data （available at： http：//www.fishgenome.org/ polarbank/） from Notothenia angustata, a species that lives in temperate near-shore water of southern New Zealand. Results from the com- parison were consistent with results inferred from phenotype differences and also with our previously published Gene Ontology-based method. The Pro-Diffprogram and operation manual can be downloaded from： http：//www.fishgenome.org/download/Prodiff.rar.