In many clinical trials, the prediction of patient outcome following therapy requires the analysis of two or more small groups of responders having a large number of simultaneously measured covariates, some of whose values may be absent. Prediction of individual outcomes in these groups is a severe statistical problem. This has motivated us to develop a suitable approach for inference from such limited data. A new statistically-oriented prediction method, called optimized independent segment voting (OISV), is presented for constructing a class-membership prediction function for such data sets. This "voting" prediction function is constructed based on the most informative and robust discrete segments of all covariate ranges, which are thus discretized.

How can one make a prediction for an individual outcome, given ill-structured and multi-scaled data about a small group of other similar individuals? To answer this question, we develop a robust combinatorial-statistical optimization method for doing pattern recognition based on such data. This method is based on the simulation of recognition processes in the immune system. Our method demonstrates higher robustness and predictive power compared to the classification and regression trees (CART) method in predicting the outcome of immunotherapy for superficial bladder cancer based on immunological measurements.

Spiral hashing is an interesting extendible hashing method. Only one approximate analysis of its performance was given, which was based on a heuristic approach and did not specify the error. In this paper, we present a discrete analysis of the expected search cost and the expected amortized rehashing cost in spiral hashing for which a big-O term can be given to describe the error

When a program is executed in a caching environment, the caching algorithm can be modeled by an associated finite-state automaton. It is assumed that the finite automaton will reach a steady state after processing a long string. By considering the finite automaton, a formula is obtained for the expected page-fault rate in terms of the steady-state probabilities of the automaton. It is possible to derive the steady-state probabilities for the least-recently-used (LRU) algorithm with order-0 and order-1 programs based on a method that describes the page reference strings as regular expressions. The steady-state behavior for caching algorithms with order-1 programs has never been reported before. This analysis method is then applied to obtain an analysis of the caching behavior of a practical storage-and-retrieval algorithm.

Abstract Varieties of files are reviewed and the notion of a file space is discussed. A number of properties which characterize file spaces are listed. This modest theoretical framework has a clarifying influence on the subject of files. Certain file‐system primitives are introduced which are similar to those in most commonly‐known operating systems, and then these primitives are used in constructing a collection of routines for reading and writing extedible arrays. Extendible arrays are based on fixed page size randomly‐accessible files and they are a useful universal input/output (I/O) structure for many different storage and retrieval schemes. The routines for managing such files are similar to the mechanisms used in handling a paged address space in a virtual memory system. The contribution of this paper lies in the detailed implementation which is presented; these routines are well understood in principle, but tricky to construct in practice. Finally, these routines are used in a practical example which provides for managing a variable blocksize free space list within a extendible array.

The object of this paper is to survey various hashing functions, to present a brief history of hashing schemes and their development, and to give an exhaustive bibliography on hashing and hash table storage and retrieval methods.

A storage and retrieval scheme which places items to be stored at the nodes of a binary tree is discussed. The tree is always balanced in a certain sense thus insuring that no excessively long search paths can exist. In addition to presenting the storage and retrieval algorithms, the deletion problem is also solved. The programming approaches involved yield a non-trivial case study of list-processing techniques. Finally, a cost analysis is given.