Comparison of algorithms for error prediction in manufacturing with automl and a cost-based metric

Journal of Intelligent Manufacturing - In order to manufacture products at low cost, machine learning (ML) is increasingly used in production, especially in high wage countries. Therefore, we...     

Approaching the accuracy–cost conflict in embedded classification system design

Smart embedded systems often run sophisticated pattern recognition algorithms and are found in many areas like automotive, sports and medicine. The developer of such a system is often confronted with the accuracy–cost conflict as the resulting system should be as accurate as possible while being able to run on resource constraint hardware. This article introduces a method to support the solution of this design conflict with accuracy–cost reports. These reports compare classification systems regarding their classification rate (accuracy) and the mathematical operations and parameters of the working phase (cost). Our method is used to deduce the specific cost of various popular pattern recognition algorithms and to derive the overall cost of a classification system. We also show how our analysis can be used to estimate the computational cost for specific hardware architectures. A software toolbox to create accuracy–cost reports was implemented to facilitate the automatic classification system comparison with the presented methodology. The software is available for download and as supplementary material. We performed different experiments on synthetic and real-world data to underline the value of this analysis. Accurate and computationally cheap classification systems were easily identified. We were even able to find a better implementation candidate in an existing embedded classification problem. This work is the first step towards a comprehensive support tool for the design of embedded classification systems.     

Advancing a Gateway Infrastructure for Wind Turbine Data Analysis

The increasing amount of data produced in many scientific and engineering domains creates as many new challenges for an efficient data analysis, as possibilities for its application. In this paper, we present one of the use cases of the project VAVID, namely the condition monitoring of sensor information from wind turbines, and how a data gateway can help to increase the usability and security of the proposed system. Starting by briefly introducing the project, the paper presents the problem of handling and processing large amount of sensor data using existing tools in the context of wind turbines. It goes on to describe the innovative approach used in VAVID to meet this challenge, covering the main goals, numerical methods used for analysis, the storage concept, and the architectural design. It concludes by offering a rational for the use of a data gateway as the main entry point to the system and how this is being implemented in VAVID.     

Determination of tacticity in polypropylene by FTIR with multivariate calibration

A method for determination of tacticity in polypropylene (PP) using FTIR associated with multivariate analysis is presented. Blends of PP with known tacticity were prepared with isotactic, syndiotactic, and atactic polymer and analyzed by ¹³C‐NMR. The FTIR spectra were recorded and processed through principal components regression (PCR) and partial least‐squares regression (PLS), using information from several different portions of the spectra. The method was compared with the classical methods of tacticity determination by FTIR based on the intensities of the bands at 998 cm^?1 (isotactic), 868 cm^?1 (syndiotactic), and 975 cm^?1 (internal standard), which are known to be dependent on the crystallinity of the polymer and, thus, affected by temperature and sample preparation. The models obtained with multivariate calibration, both with PCR and PLS, gave prediction errors up to fivefold smaller than that of the classical methods, and were also shown not to be heavily dependent on the bands that are affected by the crystallinity of the polymer, but rather on the methyl and methylene bendings at 1375 and 1462 cm^?1. © 2002 Wiley Periodicals, Inc. J Appl Polym Sci 85: 734–745, 2002     

Comparison of the chemical properties of iron and cobalt porphyrins and corrins

Density functional calculations have been used to compare various geometric, electronic and functional properties of iron and cobalt porphyrin (Por) and corrin (Cor) species. The investigation is focussed on octahedral M(II/III) complexes (where M is the metal) with two axial imidazole ligands (as a model of b and c type cytochromes) or with one imidazole and one methyl ligand (as a model of methylcobalamin). However, we have also studied some five-coordinate M(II) complexes with an imidazole ligand and four-coordinate M(I/II) complexes without any axial ligands as models of other intermediates in the reaction cycle of coenzyme B12. The central cavity of the corrin ring is smaller than that of porphine. We show that the cavity of corrin is close to ideal for low-spin Co(III), Co(II), and Co(I) with the axial ligands encountered in biology, whereas the cavity in porphine is better suited for intermediate-spin states. Therefore, the low-spin state of Co is strongly favoured in complexes with corrins, whereas there is a small energy difference between the various spin states in iron porphyrin species. There are no clear differences for the reduction potentials of the octahedral complexes, but [Co(I)Cor] is more easily formed (by at least 40 kJ mole(-1)) than [Fe(I)Por]. Cobalt and corrin form a strong Cobond;C bond that is more stable against hydrolysis than iron and porphine. Finally, Fe(II/III) gives a much lower reorganisation energy than Co(II/III); this is owing to the occupied d(z2) orbital in Co(II). Altogether, these results give some clues about how nature has chosen the tetrapyrrole rings and their central metal ion.     

On the syntheses of branched saturated fatty acids

In order to investigate the relation between the structure of mono-branched saturated fatty acids and their physical properties, a three-factor central composite design was constructed. For this purpose sixteen different fatty acids were prepared. The synthetic strategy was to use if possible, a few common starting materials for the preparation of most of the acids. Thus alkylation and hydrolysis of oxazolines were used for the preparation of 2-butylhexanoic acid, 2-methyloctadecanoic acid, 2-hexadecyloctadecanoic acid and 2-pentyloctadecanoic acid. A number of acids were prepared from thiophene derivatives followed by desulfurization with Raney-Nickel alloy under alkaline conditions. Thus, starting from 3-ethylthiophene, 4-ethyl-2-thiophenecarboxylic acid and 4-ethyl-2-methyl-5-thiophenecarboxylic acid were prepared, which upon desulfurization gave the desired 4-methylhexanoic acid. From 3-bromo-2-methylthiophene, 3-ethyl-2-methyl-5-thiophenecarboxylic acid was preparedvia 3-acetyl-2-methylthiophene and 3-ethyl-2-methylthiophene. Desulfurization gave 4-ethylhexanoic acid. Another approach started with 2-acylthiophenes, which were reacted with Grignard reagent to yield the appropriate olefins. By metalation and reaction with carbon dioxide, these were transformed to the corresponding 2-thiophenecarboxylic acids. Upon desulfurization, the desired fatty acids were obtained. In this way 6-propyldecanoic acid, 6-hexyldodecanoic acid, 6-methyldodecanoic acid and 6-pentylpentadecanoic acid were prepared. The remaining four acids were prepared from some of the branched acids described above through Kolbe reactions of dioic acids. Thus 16-methyloctadecanoic acid and 10-methyl-dodecanoic acid were obtained from 4-methylhexanoic acid, 16-ethyloctadecanoic acid from 4-ethylhexanoic acid and 9-pentyloctadecanoic acid from 6-pentylpentadecanoic acid.     

Tetramethylguanidine catalyzed transesterification of fats and oils: A new method for rapid determination of their composition

A new method for the rapid preparation of methyl esters from vegetable oils and fats, using tetramethylguanidine as catalyst, has been developed. This method is compared with the traditional method (Ce 2–66, using 20% BF₃ in methanol) and that proposed by Hartman and Lago (using NH₄Cl/H₂SO₄ in methanol). It is shown that the new method produces the methyl esters in quantitative yields and has certain advantages, as it is simpler, cheaper and does not result in isomerization of the fatty acids.