Software cost estimation with incomplete data

The construction of software cost estimation models remains an active topic of research. The basic premise of cost modeling is that a historical database of software project cost data can be used to develop a quantitative model to predict the cost of future projects. One of the difficulties faced by workers in this area is that many of these historical databases contain substantial amounts of missing data. Thus far, the common practice has been to ignore observations with missing data. In principle, such a practice can lead to gross biases and may be detrimental to the accuracy of cost estimation models. We describe an extensive simulation where we evaluate different techniques for dealing with missing data in the context of software cost modeling. Three techniques are evaluated: listwise deletion, mean imputation, and eight different types of hot-deck imputation. Our results indicate that all the missing data techniques perform well with small biases and high precision. This suggests that the simplest technique, listwise deletion, is a reasonable choice. However, this will not necessarily provide the best performance. Consistent best performance (minimal bias and highest precision) can be obtained by using hot-deck imputation with Euclidean distance and a z-score standardization     

On the generation of test patterns for multiple faults

This article presents a new method to generate test patterns for multiple stuck-at faults in combinational circuits. We assume the presence of all multiple faults of all multiplicities and we do not resort to their explicit enumeration: the target fault is a single component of possibly several multiple faults. New line and gate models are introduced to handle multiple fault effect propagation through the circuits. The method tries to generate test conditions that propagate the effect of the target fault to primary outputs. When these conditions are fulfilled, the input vector is a test for the target fault and it is guaranteed that all multiple faults of all multiplicities containing the target fault as component are also detected. The method uses similar techniques to those in FAN and SOCRATES algorithms to guide the search part of the algorithm, and includes several new heuristics to enhance the performance and fault detection capability. Experiments performed on the ISCAS'85 benchmark circuits show that test sets for multiple faults can be generated with high fault coverage and a reasonable increase in cost over test generation for single stuck-at faults.     

A Survey of Engineering Education Outside the United States: Implications for the Ideal Engineering Program

Undergraduate engineering programs in fourteen countries are surveyed to determine the status of engineering education outside the United States and to analyze the implications for the ideal engineering program of the future here in the United States. The items surveyed include the number of years required to obtain an engineering degree, title of the engineering degree, high school preparation for engineering programs, cost of education, completion rates for engineering degrees, entrance requirements, mathematical requirements, and support for laboratory work. The most notable aspects of engineering education outside the United States are:

• ? Stronger high school preparation and requirements in mathematics and science
• ? Longer period of time required for an engineering degree
• ? Advanced level of mathematics required in engineering programs
• ? Low cost of education

It appears that in most countries engineering education is more intense and rigorous than in the United States. In formulating the ideal engineering program of the future we must be sensitive to competing programs in the rest of the world.     

Effect of heat treatment on the structural and optical properties of CuInTe2 thin films

The CuInTe2 thin films were prepared by thermal vacuum evaporation of the bulk compound. The structural and optical properties in the temperature range 300–47 K of thin films grown on glass substrates and annealed in vacuum were studied. The films were investigated by X-ray diffraction and electron microscope techniques. The calculated lattice constants for CuInTe₂ powder were found to bea=0.619 nm andc –1,234 nm. From the reflection and transmission data, the optical constants, refractive indexn, absorption index,k, and the absorption coefficient, , werw computed. The optical energy gap was determined for CuInTe₂ thin films heat treated at different temparatures for different periods of time. It was found thatE _g increases with both increasing temperature and time of annealing.     

Explanation-Based Learning for Diagnosis

We present explanation-based learning (EBL) methods aimed at improving the performance of diagnosis systems integrating associational and model-based components. We consider multiple-fault model-based diagnosis (MBD) systems and describe two learning architectures. One, EBLIA, is a method for learning in advance. The other, EBL(p), is a method for learning while doing. EBLIA precompiles models into associations and relies only on the associations during diagnosis. EBL(p) performs compilation during diagnosis whenever reliance on previously learned associational rules results in unsatisfactory performance—as defined by a given performance threshold p. We present results of empirical studies comparing MBD without learning versus EBLIA and EBL(p). The main conclusions are as follows. EBLIA is superior when it is feasible, but it is not feasible for large devices. EBL(p) can speed-up MBD and scale-up to larger devices in situations where perfect accuracy is not required.     

Computational complexity of determining resource loops inre-entrant flow lines

This paper presents a comparison study of the computational complexity of the general job shop protocol and the more structured flow line protocol in a flexible manufacturing system. It is shown that the representative problem of finding resource invariants is NP-complete in the case of the job shop, while in the flow line case it admits a closed form solution. The importance of correctly selecting part flow and job routing protocols in flexible manufacturing systems to reduce complexity is thereby conclusively demonstrated     

Ferrimagnetic resonance and low temperature transition in Ni- and Co-substituted powder magnetite

The effects of nickel and cobalt substitution on ferrimagnetic resonance in high-purity magnetite powder have been investigated. The low-temperature transition temperature and damping parameter, , are found to decrease with increasing solute additions. The linewidth, H, varies in proportion to the anisotropy field of the ferrites, narrowing with increasing nickel content but broadening with increasing cobalt content. Theg factor increases with the dopant concentration, passing through a maximum around room temperature. Decreasing the deviation from stoichiometry by appropriate annealing decreases the transition temperature shift but slightly increases H, g and .     

Tissue growth patterns in the carcasses of water buffalo and Friesian crossbred cattle-part 1: Individual muscles and anatomical muscle groups

The left sides of the carcasses of twelve Egyptian buffalo, nine half Friesian (♂ Friesian x ♀ Egyptian Baladi) and nine three-quarter Friesian (♂ x ♀ half-Friesian) bulls, serially slaughtered between 161 and 560 kg for buffaloes and between 176 and 448 kg for cattle, were used to describe the growth and distribution of individual muscles and muscle groups.

Genotype-group differences in the relative rate of growth were not significant in 85 of 89 muscles and in 8 of 9 muscle groups. The pooled within-group growth coefficients of individual muscles revealed different increasing growth gradients, i.e. disto-proximal in both limbs, mediolateral in the proximal pelvic limb, from elbow flexors to shoulder flexors in the proximal thoracic limb, caudo-cranial in the trunk and from hypaxial to epaxial muscles around the spinal column. The relative growth was lowest in the distal part of the limbs and highest in the anterior part of the carcass (SMG 7 and 8), with the loin and abdominal muscle groups growing at a rate similar to that of the total muscle. The neck and thorax muscle group grew at a higher rate in Friesian crosses than in buffaloes.

Genotype-group differences in the weight of each muscle relative to total side muscle (TSM) were significant in 50 of 89 muscles. Of the 50 muscles (constituting around 59% of TSM), 24 (approximately 44% of TSM) were from the expensive muscle groups. As compared with the most different Friesian cross (base = 100) at equal TSM, buffaloes had higher weight of muscle in the hindlimb (107·5% proximally; 106·3% distally) and forelimb (120·0% proximally: 104·6% distally) and significantly less weight of the muscle groups forming the abdominal wall (79·8%) and connecting the forelimb to the thorax (89·1%) and to the neck (90·7%). The weight of the combined expensive groups was significantly greater in buffaloes than in cattle (maximum difference = 1·8 kg) with a tendency for buffaloes to have relatively less of the tender muscles surrounding the spinal column.     

Intracellular localization of samarium in the lactating mammary gland cells: ultrastructural and microanalytical study

The frequent use of some rare earths in the medical and industrial domains make us worry about their intracellular behavior into the body. Reason for which we have investigated the subcellular localization of one of these elements, the samarium, in the mammary gland of lactating female wistar rats using two very sensitive methods of observation and microanalysis, the transmission electron microscopy and the secondary ion mass spectrometry. The ultrastructural study showed the presence of electron dense deposits in the lactating mammary glandular epithelial cell lysosomes of the samarium-treated rats, but no loaded lysosomes were observed in those of control rats. The microanalytical study allowed both the identification of the chemical species present in those deposits as samarium isotopes ((152) Sm(+)) and the cartography of its distribution. Our results confirm the previous ones showing that lysosomes of the glandular epithelial cells are the site of the intracellular concentration of foreign elements such as gallium. The intralysosomal deposits observed in the mammary glandular cells of the samarium-treated rats are similar in their form and density to those observed with the same element in other varieties of cells, such as liver, bone marrow, and spleen cells. Our ultrastructural and microanalytical results and those obtained in previous studies allow deducing that the intralysosomal deposits are very probably composed of an insoluble samarium phosphate salt.     

Assessment of sediment quality and pore water ecotoxicity in Kebir Rhumel basin (NE-Algeria): a combined approach

The objectives of this study are to use different approaches to assess the current pollution status in the wadis of the Kebir Rhumel basin. First, sediment trace metal contents were measured by flame atomic absorption spectroscopy. Then, sediment quality was assessed on the basis of contamination assessment indexes such as: Geoaccumulation Index (Igeo), Contamination factor (C(f)), Contamination degree (C(d)), Sediment Pollution Index (SPI) and SEQ guidelines (Consensus Sediment Quality Guidelines). In addition, several toxicity tests (Daphnia magna mobility inhibition acute test-48 h, Aliivibrio fischeri luminescence inhibition acute test - 15/30 mn and Pseudokirchneriella subcapitata growth inhibition chronic test - 72 h) were conducted to assess sediment pore water ecotoxicity. Trace metal concentrations followed the order: Mn > Zn > Pb > Cr > Cu > Ni > Co > Cd. Indexes used indicate varying degrees of sediment quality. Igeo, C(f), C(d) and SPI reveal a polymetallic contamination dominated by two or more elements in which Cd, Cu and Pb are of greatest concern. SEQ guidelines showed that biological effects on fauna would likely be observed occasionally and/or frequently for Cd, Cr, Cu, Pb and Zn contents. Test organisms exposed to sediment pore water showed that the algal P. subcapitata test was more sensitive than the D. magna and A. fischeri tests. Hence, algal growth inhibition proved to be the most sensitive response to contaminants present in sediment extracts but a significant relationship with trace metal contents was not demonstrated.