Automated bug assignment: Ensemble-based machine learning in large scale industrial contexts

Bug report assignment is an important part of software maintenance. In particular, incorrect assignments of bug reports to development teams can be very expensive in large software development projects. Several studies propose automating bug assignment techniques using machine learning in open source software contexts, but no study exists for large-scale proprietary projects in industry. The goal of this study is to evaluate automated bug assignment techniques that are based on machine learning classification. In particular, we study the state-of-the-art ensemble learner Stacked Generalization (SG) that combines several classifiers. We collect more than 50,000 bug reports from five development projects from two companies in different domains. We implement automated bug assignment and evaluate the performance in a set of controlled experiments. We show that SG scales to large scale industrial application and that it outperforms the use of individual classifiers for bug assignment, reaching prediction accuracies from 50 % to 89 % when large training sets are used. In addition, we show how old training data can decrease the prediction accuracy of bug assignment. We advice industry to use SG for bug assignment in proprietary contexts, using at least 2,000 bug reports for training. Finally, we highlight the importance of not solely relying on results from cross-validation when evaluating automated bug assignment.     

Effect of calcination atmosphere and temperature on γ-Al2O3 supported cobalt Fischer-Tropsch catalysts

The effect of calcination temperature and atmosphere on the properties of γ-Al₂O₃ supported cobalt Fischer-Tropsch catalysts has been investigated. One common precursor for all the catalysts was prepared by incipient wetness impregnation of γ-Al₂O₃ with an aqueous solution of cobalt nitrate hexahydrate. It was subjected to four different calcination atmospheres (air/50% steam: 30 mL/min, air: 30 mL/min, air: 50 mL/min, N₂: 30 mL/min) and eight different calcination temperatures (range: 473–723 K), making the total number of samples 32. Both the post calcination nitrogen content and the cobalt dispersion were measured. The results demonstrated that in order to maximise the cobalt dispersion, it is necessary to use low calcination temperatures and remove the precursor decomposition products (NO, NO₂, H₂O) efficiently. The Fischer-Tropsch synthesis performance of two catalysts calcined at the same temperature, but at different air flow rates was evaluated. No significant effect of the air flow rate was found on the turnover frequency or C₅₊ selectivity, but a high flow rate resulted in 30% higher activity per gram catalyst.     

Minimizing costs in round robin tournaments with place constraints

In this paper we consider a general sports league scheduling problem and propose solution algorithms for it. The objective is to find a feasible schedule for a round robin tournament with minimum number of breaks and minimum total costs where additionally place constraints are taken into account. We present a “first-break, then-schedule” approach which uses an enumerative procedure to generate home-away patterns and integer programming for finding corresponding schedules. Computational results are presented for leagues with up to 14 teams.     

Time‐resolved small‐angle X‐ray scattering study of void fraction evolution in high‐density polyethylene during stress unloading and strain recovery

By means of time‐resolved small‐angle X‐ray scattering, we developed an analysis methodology to assess the void volume fraction ?_v in high‐density polyethylene (HDPE) during tensile testing. The specimens were first drawn up to different imposed strains, and subsequently were subjected to stress unloading and strain recovery stages. During the loading stage, ?_v progressively increased with the strain level, starting from a well‐defined onset strain prior to the yield point. In particular, ?_v reached a maximum of 8.75 vol% for a strain of 12.5% in the case of a HDPE grade with a molecular weight of 105 000 g mol^?1. Stress unloading and strain recovery caused a decrease in ?_v attained at the end of the loading stage. For a HDPE grade with a molecular weight of 55 000 g mol^?1, ?_v was more important during the loading stage and the decrease in ?_v was less marked during the stress unloading stage when compared to the HDPE with molecular weight of 105 000 g mol^?1. The residual and reversible components of void volume fraction were revealed. © 2015 Society of Chemical Industry     

Methane yield of oat husks

The biogas yield of solid manure from dairy cattle depends on its quality and the proportion of excreta and organic litter material contained within. The biogas yield of both faeces and straw is available in literature. Straw is a common litter material of mixed farms. However, straw is scarcely available on dairy farms. Oat husks are appropriate to replace or supplement straw for use as litter material. In this study, the actual methane yield and the total methane potential of oat husks were determined. Based on an optimized test with ground oat husks, the total methane potential resulted from regression and extrapolation of the experimental data. The total methane potential was determined with 242 L_N CH₄ kg⁻¹ VS added. Additionally, the actual methane yield over retention time at a digestion temperature of 37 °C was determined, using untreated oat husks. For 42 days of retention, the methane yield was 202 L_N CH₄ kg⁻¹ VS added at 52% CH₄ content. Results indicate that the methane yield of oat husks reaches the same level as that of straw. The total methane potential is not higher, but digestion of oat husks may proceed faster. Verification of the laboratory results on-farm revealed that the contribution of oat husks to overall methane production of a prototype biogas plant for solid manure might reach up to 80%.     

Acidified Nitrite Contributes to the Antitumor Effect of Cold Atmospheric Plasma on Melanoma Cells

Cold atmospheric plasma (CAP) is partially ionized gas near room temperature with previously reported antitumor effects. Despite extensive research and growing interest in this technology, active components and molecular mechanisms of CAP are not fully understood to date. We used Raman spectroscopy and colorimetric assays to determine elevated nitrite and nitrate levels after treatment with a MiniFlatPlaster CAP device. Previously, we demonstrated CAP-induced acidification. Cellular effects of nitrite and strong extracellular acidification were assessed using live-cell imaging of intracellular Ca²⁺ levels, cell viability analysis as well as quantification of p21 and DNA damage. We further characterized these observations by analyzing established molecular effects of CAP treatment. A synergistic effect of nitrite and acidification was found, leading to strong cytotoxicity in melanoma cells. Interestingly, protein nitration and membrane damage were absent after treatment with acidified nitrite, thereby challenging their contribution to CAP-induced cytotoxicity. Further, phosphorylation of ERK1/2 was increased after treatment with both acidified nitrite and indirect CAP. This study characterizes the impact of acidified nitrite on melanoma cells and supports the importance of RNS during CAP treatment. Further, it defines and evaluates important molecular mechanisms that are involved in the cancer cell response to CAP.