Automatic Selection of Classification Algorithms for Non-Experts Using Meta-Features

With the arrival of big-data society, methods for classifying real-world problems have attracted much attention for researchers and developers in various fields. In recent years, much effort has been devoted for improving performances of classification algorithms by adding functions or modifying their weaknesses. However, since a large variety of classification algorithms has been available, it is difficult for non-experts to find classification algorithms that achieve good results on a given data set. Therefore, if there is a system which automatically selects the best classification algorithm for a given data set, non-experts would receive various benefits such as saving time and effort. This paper presents a system of predicting the best possible classification algorithm for a given data set with respect to the accuracy. To the best of our knowledge, this is the first approach focused on predicting the best one. The main target users of the proposed system are non-experts who do not have knowledge and experience in data mining. The proposed system utilizes useful meta-features selected from existing recta-features to increase the performance of the prediction. The feature selection is conducted by a wrapper approach with the genetic search algorithm. In the proposed system, K-nearest neighbor algorithm is used to learn the selectedmeta-features and build a classification model for predicting future data. Experiments using 58 real-world data sets show that the proposed system predicted the best classification algorithm with 60.34% accuracy from the top five in 30 classification algorithms.     

Price setting in forward-looking customer markets

If consumers form habits in individual goods, firms face a time-inconsistency problem. Low prices in the future help attract customers in the present. Firms, therefore, have an incentive to promise low prices in the future, but price gouge when the future arrives. In this setting, firms benefit from “committing to a sticky price.” If consumers have incomplete information about costs and demand, the firm-preferred equilibrium has the firm price at or below a “price cap.” The model therefore provides an explanation for the simultaneous existence of a rigid regular price and frequent “sales”.     

Productivity versus elasticity: a normalized constant elasticity of substitution production function applied to historical Soviet data

This study tests the extensive growth hypothesis, which would attribute the Soviet economic slowdown to low elasticity of substitution and over-investment rather than deteriorating productivity growth. To circumvent the low availability of data, widely applicable new methods for estimating productivity and elasticity of substitution were developed based on the normalized constant elasticity of substitution (CES) production function. One is an extended version of the Solow residual calculation, and the other is a direct estimation of the time-varying parameter normalized CES production function. The application of the methods to the Soviet data showed a decreasing trend in productivity growth and a low elasticity of substitution of approximately 0.25. The results neither reject nor support the extensive growth hypothesis because of uncertainty in the interpretation of the low elasticity of substitution. The results suggest that it is reasonable to assert that both deteriorating productivity growth and low elasticity of substitution caused the Soviet economic slowdown. Further empirical studies on productivity and elasticity of substitution in the Soviet and other economies are necessary to fully understand relations between productivity growth, elasticity of substitution and economic development.     

A two-stage model for the design of supply chain networks

This research aims to develop an analytical model of the supplier selection process in designing a supply chain network. The constraints on the capacity of each potential supplier are considered in the process. The assumed objective of the supply chain is to minimize the level of customer dissatisfaction, which is evaluated by two performance criteria: (i) price and (ii) delivery lead time. The overall model operates at two levels of decision-making: the operational level and the chain level. The operational level concerns decisions related to optimizing the manufacturing and logistical activities of each potential supplier in order to meet the customer's requirements. At the chain level, all the bids from potential suppliers are evaluated and the final configuration of the supply chain is determined. The structure of the chain depends on the product specifications and on the customer's order size. An optimal solution in terms of the models for the two levels can be obtained by using a mixed-integer programming technique.     

Is living in African cities expensive?

Although several studies have examined why overall price levels are higher in richer countries, little is known about whether there is a similar relationship at the urban and city level across countries. This paper compares the price levels of cities in Sub-Saharan Africa with those of other regions by analyzing price information collected for the purpose of calculating official purchasing power parities. The approach of the paper is to readjust the calculated price levels from national to urban levels, using known price-level ratios between those areas. The results indicate that African cities are relatively more expensive, despite having lower income levels. The price levels of goods and services consumed by households are 25 to 28 percent higher in Sub-Saharan Africa than in other low- and middle-income countries, relative to their income levels. Such high costs of living could constrain livelihood of low-income urban residents, as well as the development of Africa’s urban economies.     

R&D boundaries of the firm: An estimation of the double-hurdle model on commissioned R&D,joint R&D,and licensing in Japan

This paper studies the firm’s decisions on in-house R&D and its procurement from outside through commissioned R&D, joint R&D, and technology acquisitions (i.e., licensing-in). Using the data about 14,000 manufacturing firms in Japan, we estimate a modified double-hurdle model in which the first hurdle determines whether the firm should perform any R&D at all and the second hurdle determines whether (and how much) it should perform each mode of procured R&D. The results generally support the two major theories—the transaction cost theory and the capability theory. The estimated positive effects of firm size, in-house R&D intensity, diversification, and vertical integration support the hypothesis that capability is needed for procured R&D, while the estimated positive effect of the index of appropriability by patents supports the hypothesis that this appropriability reduces transaction costs. In addition, we found that information flow from scientific sources and that from transaction-based sources affect the three modes of procured R&D differently.     

The housing bubble and a new approach to accounting for housing in a CPI

Over the course of the recent house price bubble in the United States, the price of homes rose rapidly from 1999 Q4 to 2005 Q4 (11.3% annually as measured by the Case-Shiller index, and 8.4% annually as measured by the Federal Housing Financing Agency) but slowly as measured by owner equivalent rents (3.4%), so measured core inflation remained relatively docile during this period, since only rents are used to measure inflation for housing services in the United States. Over the last several decades, the US Bureau of Labor Statistics (BLS) has experimented with both rental equivalence and user cost approaches for accounting for owner occupied housing (OOH) services in the CPI. We explain the basics of these approaches, and outline the BLS experiences with using them. This assessment leads us to conclude that the time has come to try a new approach: the opportunity cost approach. We argue this approach has advantages over both the conventional rental equivalence and user cost approaches, though it embeds components of the measures for both those approaches and builds solidly on the research of Verbrugge and others at the BLS. Also, we take up empirical issues that must be faced regardless of which of the approaches discussed is adopted. We explain how the repeat-sales and various hedonic regression methods can be placed in a common framework, thereby facilitating understanding of the properties of and the tradeoffs between the methods. We also consider measurement complications that arise because the land and structure components of properties depreciate at different rates.     

Stable market structures from merger activities in mixed oligopoly with asymmetric costs

This paper examines endogenous merger formations in a mixed oligopoly. Applying the core as a solution concept, we analyze which market structure(s) remain(s) stable when three firms—two symmetric private firms and one inefficient public firm—are allowed to merge with each other in a mixed Cournot industry. We show that according to the value of the marginal cost of the public firm, there always exists a pair of share ratios of the owners of both the (pre-merged) public firm and the (pre-merged) private firm such that the market structure with the merger between the public firm and one private firm belongs to the core. When the initial market structure is a mixed triopoly, it can only be blocked when one public firm and one private firm merge. Furthermore, we conduct a similar analysis in a general mixed oligopoly with one public firm and n private firms.      

The price of education and inequality

Constructing a simple model that includes the price of education, this paper shows that the educational expenditure of rich households could prevent poor households from escaping poverty. The paper offers an explanation for persistent inequality.