Improved customer choice predictions using ensemble methods

In this paper various ensemble learning methods from machine learning and statistics are considered and applied to the customer choice modeling problem. The application of ensemble learning usually improves the prediction quality of flexible models like decision trees and thus leads to improved predictions. We give experimental results for two real-life marketing datasets using decision trees, ensemble versions of decision trees and the logistic regression model, which is a standard approach for this problem. The ensemble models are found to improve upon individual decision trees and outperform logistic regression.     

多项选择题测验信度的一个估计方法及其应用

在多项选择题测验分数等于真实分数和猜测分数之和的假设模型下，本文得到了一个多项选择题测验信度的理论公式，并由此给出了测验信度的估计方法。最后，通过两个例子，说明了本文提出的方法在测验信度分析中的应用，并将这种方法与教育测量中常用的库德——理查逊方法（Ｋｕｄｅｒ－Ｒｉｃｈａｒｄｓｏｎ）进行了比较     

On the choice of active space orbitals in MCSCF calculations

We describe a procedure which may be used to aid selection of the active space in multiconfigurational self-consistent field (MCSCF) calculations for general chemical systems. Starting from a restricted Hartree-Fock calculation, we define a hierarchy of interacting virtual orbitals for every occupied orbital. The most strongly interacting orbitals are then taken to constitute the active space in a configuration interaction (CI) calculation. The natural orbital occupation numbers obtained from the CI calculation are then used to choose the active space to be used in a subsequent MCSCF calculation. We illustrate our method on a number of systems (Li₂, B₂, C₂, carbonyl oxide and the transition state for oxidation of H₂S by dioxirane). In all these cases, ‘intuitive’ active spaces are inadequate, as are active spaces derived from the natural orbitals of unrestricted Hartree-Fock calculations.     

Choiceless polynomial time

Turing machines define polynomial time (PTime) on strings but cannot deal with structures like graphs directly, and there is no known, easily computable string encoding of isomorphism classes of structures. Is there a computation model whose machines do not distinguish between isomorphic structures and compute exactly PTime properties? This question can be recast as follows: Does there exist a logic that captures polynomial time (without presuming the presence of a linear order)? Earlier, one of us conjectured a negative answer. The problem motivated a quest for stronger and stronger PTime logics. All these logics avoid arbitrary choice. Here we attempt to capture the choiceless fragment of PTime. Our computation model is a version of abstract state machines (formerly called evolving algebras). The idea is to replace arbitrary choice with parallel execution. The resulting logic expresses all properties expressible in any other PTime logic in the literature. A more difficult theorem shows that the logic does not capture all of PTime.     

Extreme point axioms for closure spaces

A pair (X,τ) of a finite set X and a closure operator τ:2^X→2^X is called a closure space. The class of closure spaces includes matroids as well as antimatroids. Associated with a closure space (X,τ), the extreme point operator ex:2^X→2^X is defined as ex(A)={p|p∈A,p∉τ(A-{p})}. We give characterizations of extreme point operators of closure spaces, matroids and antimatroids, respectively.     

<Emphasis Type="Bold">R</Emphasis><Emphasis Type="SmallCaps">UBIS</Emphasis>: a bipolar-valued outranking method for the choice problem

The main concern of this article is to present the R UBIS method for tackling the choice problem in the context of multiple criteria decision aiding. Its genuine purpose is to help a decision maker to determine a single best decision alternative. Methodologically we focus on pairwise comparisons of these alternatives which lead to the concept of bipolar-valued outranking digraph. The work is centred around a set of five pragmatic principles which are required in the context of a progressive decision aiding methodology. Their thorough study and implementation in the outranking digraph lead us to define a choice recommendation as an extension of the classical digraph kernel concept.      

The non-Markovian nature of nested logit choice

Discrete choice models are widely used for understanding how customers choose between a variety of substitutable goods. We investigate the relationship between two well studied choice models, the Nested Logit (NL) model and the Markov choice model. Both models generalize the classic Multinomial Logit model and admit tractable algorithms for assortment optimization. Previous evidence indicates that the NL model may be well approximated by, or be a special case of, the Markov model. We establish that the Nested Logit model, in general, cannot be represented by a Markov model. Further, we show that there exists a family of instances of the NL model where the choice probabilities cannot be approximated to within a constant error by any Markov choice model.     

Application of fuzzy soft set in decision making problems based on grey theory

There are many uncertain problems in practical production and life which need decisions made with soft sets and fuzzy soft sets. However, the basis of evaluation of the decision method is single and simple, the same decision problem can obtain different results from using a different evaluation basis. In this paper, in order to obtain the right result, we discuss fuzzy soft set decision problems. A new algorithm based on grey relational analysis is presented. The evaluation bases of the new algorithm are multiple. There is more information in a decision result based on multiple evaluation bases, which is more easily accepted and logical to one’s thinking. For the two cases examined, the results show that the new algorithm is efficient for solving decision problems.     

Georgescu模糊选择函数伪传递合理性的研究

本文研究模糊选择函数伪传递合理性。对普通选择函数伪传递合理的两个充要条件进行了模糊化,讨论了它们的独立性;证明了对Georgescu模糊选择函数而言,这两个条件仍是伪传递合理的充分条件,但不再是必要条件。