Root cause prediction for failures in semiconductor industry, a genetic algorithm-machine learning approach.

Failure analysis has become an important part of guaranteeing good quality in the electronic component manufacturing process. The conclusions of a failure analysis can be used to identify a component's flaws and to better understand the mechanisms and causes of failure, allowing for the implementation of remedial steps to improve the product's quality and reliability. A failure reporting, analysis, and corrective action system is a method for organizations to report, classify, and evaluate failures, as well as plan corrective actions. These text feature datasets must first be preprocessed by Natural Language Processing techniques and converted to numeric by vectorization methods before starting the process of information extraction and building predictive models to predict failure conclusions of a given failure description. However, not all-textual information is useful for building predictive models suitable for failure analysis. Feature selection has been approached by several variable selection methods. Some of them have not been adapted for use in large data sets or are difficult to tune and others are not applicable to textual data. This article aims to develop a predictive model able to predict the failure conclusions using the discriminating features of the failure descriptions. For this, we propose to combine a Genetic Algorithm with supervised learning methods for an optimal prediction of the conclusions of failure in terms of the discriminant features of failure descriptions. Since we have an unbalanced dataset, we propose to apply an F1 score as a fitness function of supervised classification methods such as Decision Tree Classifier and Support Vector Machine. The suggested algorithms are called GA-DT and GA-SVM. Experiments on failure analysis textual datasets demonstrate the effectiveness of the proposed GA-DT method in creating a better predictive model of failure conclusion compared to using the information of the entire textual features or limited features selected by a genetic algorithm based on a SVM. Quantitative performances such as BLEU score and cosine similarity are used to compare the prediction performance of the different approaches.

Predicting Fluctuating Rates of Hospitalizations in Relation to Influenza Epidemics and Meteorological Factors.

INTRODUCTION: In France, rates of hospital admissions increase at the peaks of influenza epidemics. Predicting influenza-associated hospitalizations could help to anticipate increased hospital activity. The purpose of this study is to identify predictors of influenza epidemics through the analysis of meteorological data, and medical data provided by general practitioners. METHODS: Historical data were collected from Meteo France, the Sentinelles network and hospitals' information systems for a period of 8 years (2007-2015). First, connections between meteorological and medical data were estimated with the Pearson correlation coefficient, Principal component analysis and classification methods (Ward and k-means). Epidemic states of tested weeks were then predicted for each week during a one-year period using linear discriminant analysis. Finally, transition probabilities between epidemic states were calculated with the Markov Chain method. RESULTS: High correlations were found between influenza-associated hospitalizations and the variables: Sentinelles and emergency department admissions, and anti-correlations were found between hospitalizations and each of meteorological factors applying a time lag of: -13, -12 and -32 days respectively for temperature, absolute humidity and solar radiation. Epidemic weeks were predicted accurately with the linear discriminant analysis method; however there were many misclassifications about intermediate and non-epidemic weeks. Transition probability to an epidemic state was 100% when meteorological variables were below: 2°C, 4 g/m3 and 32 W/m2, respectively for temperature, absolute humidity and solar radiation. This probability was 0% when meteorological variables were above: 6°C, 5.8g/m3 and 74W/m2. CONCLUSION: These results confirm a good correlation between influenza-associated hospitalizations, meteorological factors and general practitioner's activity, the latter being the strongest predictor of hospital activity.

Analysis of odorous compounds at municipal landfill sites.

The aim of this investigation was to determine odorous compounds in the air over the landfill sites in France and Poland. Air samples were collected by passive and dynamic methods of preconcentration analytes and analysed by GC-MS and GC-FID. The coupling microTD-microGC-MS was also used for on-site analysis of odorous compounds. The achieved results indicated that the concentrations of odorants in the air varied and strongly depended on the sampling site. The highest concentrations were observed at the points situated near biogas wells and above the fresh waste layer. The concentrations were influenced by landfill activities such as failures of the landfill gas collection system, heavy truck traffic, machinery operations and compacting fresh waste.