Study on Data Preprocessing for Machine Learning Based on Semiconductor Manufacturing Processes.

Various data types generated in the semiconductor manufacturing process can be used to increase product yield and reduce manufacturing costs. On the other hand, the data generated during the process are collected from various sensors, resulting in diverse units and an imbalanced dataset with a bias towards the majority class. This study evaluated analysis and preprocessing methods for predicting good and defective products using machine learning to increase yield and reduce costs in semiconductor manufacturing processes. The SECOM dataset is used to achieve this, and preprocessing steps are performed, such as missing value handling, dimensionality reduction, resampling to address class imbalances, and scaling. Finally, six machine learning models were evaluated and compared using the geometric mean (GM) and other metrics to assess the combinations of preprocessing methods on imbalanced data. Unlike previous studies, this research proposes methods to reduce the number of features used in machine learning to shorten the training and prediction times. Furthermore, this study prevents data leakage during preprocessing by separating the training and test datasets before analysis and preprocessing. The results showed that applying oversampling methods, excluding KM SMOTE, achieves a more balanced class classification. The combination of SVM, ADASYN, and MaxAbs scaling showed the best performance with an accuracy and GM of 85.14% and 72.95%, respectively, outperforming all other combinations.

Spectrum of susceptibility to air quality and weather in individual children with atopic dermatitis.

BACKGROUND: Exposure to air pollution aggravates symptoms of atopic dermatitis (AD) in children in the population studies. Variability in individual patient's response from individual susceptibility is needed to be explored. OBJECTIVE: This study aimed to investigate spectrum of individual variability in the associations between AD symptoms and air quality. METHODS: We enrolled 89 children aged 0-6 years with AD (22 890 person-days). Daily manifestation of symptoms was recorded for an average of 257 days (range 100-499). Both an individual analysis using logistic regression models and an overall analysis using a generalized estimating equation were performed. RESULTS: The odds ratios of an individual ranged 0.24-8.11 for particulate matter <10 µm in diameter (PM10 ), 0.09-101.92 for nitrogen oxide (NO2 ), 0.03-44.00 for ozone (O3 ), 0.11-58.30 for sulfur dioxide (SO2 ), 0.00-15.83 for carbon monoxide (CO), 0.00-39 446.94 for temperature, and 0.03-5.18 for relative humidity, demonstrating a wide individual variability. In the overall analysis, PM10 , NO2 , SO2 , and CO had a significantly positive association, whereas temperature and relative humidity were negatively associated with AD symptoms. Air pollution was responsible for aggravation of symptoms from 24.7% (O3 ) to 39.3% (SO2 ) of AD children. Overall, 71.9% of the AD children responded to at least one or more air pollution and weather variable. CONCLUSION: Responses of AD children to air pollution and weather variable were considerably variable among individuals. An individualized model would be useful to forecast and manage AD symptoms in patients.

Exposure to phthalates and bisphenol A are associated with atopic dermatitis symptoms in children: a time-series analysis.

BACKGROUND: Despite increasing evidence on the relationship between exposure to phthalates and bisphenol A with allergies and asthma, reports on atopic dermatitis (AD) with these chemicals are few. We assessed the association between AD symptoms and the exposure to phthalates and bisphenol A and in children. METHODS: We surveyed 18 boys with AD (age 3-7 years) in a day care center in Seoul between May 2009 and April 2010. AD symptoms were recorded by using a daily symptom diary. We collected 460 series of pooled urine twice a day, in the morning and afternoon, over 230 working days and measured the concentrations of mono-2-ethyl-5-oxohexyl phthalate (5-oxo-MEHP), mono-2-ethyl-5-hydroxyhexyl phthalate (5-OH-MEHP), mono-isobutyl phthalate (MnBP) and bisphenol A glucuronide (BPAG) in the pooled urine. Logistic regression was used for statistical analysis. RESULTS: Most phthalate metabolite levels were higher in the morning than in the afternoon (p < 0.0001). There was seasonal variation in the levels of phthalates and bisphenol A metabolites. Levels of 5-OH-MEHP, MnBP, and BPAG were highest in summer (p < 0.0001). Manifestation of AD symptoms was associated with an increase in urinary levels of MnBP (adjusted odds ratio, aOR = 2.85, 95% CI: 1.12-7.26 per 1 µg/L of MnBP) and BPAG (aOR = 1.79, 95% CI: 0.91-3.52 per 1 µg/L BPAG) on the same day. The levels of MnBP and BPAG in the previous day increased AD symptoms (aOR = 2.74, 95% CI: 1.21-6.20, for 1 µg/L of MnBP and aOR = 2.01, 95% CI: 1.08-3.74 for 1 µg/L BPAG). CONCLUSION: Our results suggest that exposure to phthalates and bisphenol A is associated with aggravation of AD symptoms in children.

Effect of acid treatment on allergenicity of peanut and egg.

BACKGROUND: Many food experts have studied various treatments or processing techniques in order to develop hypoallergenic foods. In a previous study, acid treatment dramatically mitigated the allergenicity of peanut, especially Ara h 2. RESULTS: Gel electrophoresis showed that most protein bands of acid-treated peanut were not detected, but protein bands of egg white became weaker and broader by acid treatment. In immunoblotting using a rabbit antibody, the antigenicity against ovalbumin or ovomucoid in acid-treated egg white was decreased but the antigenicity against Ara h 1 or Ara h 2 in peanut treated with pH 2 acetic acid was completely undetected. The allergenicity of ovalbumin and peanut fell significantly to 1/1022 and 1/5380, respectively, when measured as IC50 in the sample treated with pH 2.0 acetic acid. CONCLUSION: This study showed that acid treatment was more effective in peanut and barely effective in ovomucoid. This may contribute to the development of hypoallergenic food and clinical management of food allergy. © 2016 Society of Chemical Industry.

Effect of the prenatal maternal environments and diets on cord blood interleukin-4 and interferon -gamma: A pilot study.

BACKGROUND AND OBJECTIVE: TThe environment of a pregnant woman can affect not only fetal growth and development, but also diseases in childhood. Neonatal cord blood cytokines are commonly used to evaluate the immune development of neonates. The purpose of this study was to evaluate the effects of the environment and diet during pregnancy on IL-4 and IFN-γ in neonatal cord blood. METHOD: A total of 111 pregnant women participated in this study from April to November 2010. Allergy history, sensitization assessed by the skin prick test, dietary intake and indoor environment were evaluated. IL-4 and IFN-γ levels were measured in the complete cord blood of neonates using real-time PCR. RESULTS: There were 54 pregnant women with allergic disease. Both IL-4 and IFN-γ levels in neonatal cord blood were higher in samples from allergic mothers than in non-allergic mothers (p<0.05). The indoor environment and nutrient intake were not different between allergic and non-allergic mothers, except regarding carpet use. When the cytokine levels were divided into quartiles, lower folate and vitamin B6 intake was associated with the highest levels of IL-4 in neonatal cord blood (p<0.05), and higher folate and vitamin B6 intake was associated with highest levels of IFN-γ in neonatal cord blood. CONCLUSIONS: In this study, a strong association between IL-4 and IFN-γ levels in cord blood and the intake of folate and vitamin B6 was found, which indicates that food intake during pregnancy might have a strong influence on IL-4 and IFN-γ levels in cord blood, to a greater extent than environmental factors.

Association of carbon monoxide levels with allergic diseases in children.

BACKGROUND: Air pollution is regarded as a potential reason for the development or aggravation of allergic diseases. OBJECTIVE: To investigate the relationship between exposure to outdoor air pollution during the first year of life and the development of allergic diseases, and to determine the association between outdoor air pollution near the residence and recent symptoms of allergic diseases in children. METHODS: A nationwide cross-sectional survey was conducted in 6- to 7-year-old children. Among them, subjects who had been born (n = 1828) or lived (n = 1894) within 2 km of the nearest air quality monitoring sites were selected. The prevalence of allergic diseases was assessed by using questionnaires. Each subject was examined by pediatricians to determine the presence of eczema. Publicly available data were used regarding the daily ambient concentrations of sulfur dioxide, nitrogen dioxide, ozone, carbon monoxide (CO), and particulate matter. RESULTS: For a 100 ppb increase in the mean daily CO level during the first year of life, the adjusted odds ratio for lifetime allergic rhinitis was 1.10 (95% confidence interval, 1.03-1.19). For a 1 ppm increase in the annual average CO level during the past 12 months, the adjusted odds ratio for the presence of atopic dermatitis on the day of the survey was 8.11 (95% confidence interval, 1.06-62.12). However, the other air pollutants showed no statistical significance. CONCLUSIONS: Higher exposure to CO during infancy increased the risk of development of allergic rhinitis. The risk of current atopic dermatitis symptoms increased in children who were exposed to higher CO levels during the past 12 months.

What Characteristics Confer Proteins the Ability to Induce Allergic Responses? IgE Epitope Mapping and Comparison of the Structure of Soybean 2S Albumins and Ara h 2.

Ara h 2, a peanut 2S albumin, is associated with severe allergic reactions, but a homologous protein, soybean 2S albumin, is not recognized as an important allergen. Structural difference between these proteins might explain this clinical discrepancy. Therefore, we mapped sequential epitopes and compared the structure of Ara h 2, Soy Al 1, and Soy Al 3 (Gly m 8) to confirm whether structural differences account for the discrepancy in clinical responses to these two proteins. Commercially synthesized peptides covering the full length of Ara h 2 and two soybean 2S albumins were analyzed by peptide microarray. Sera from 10 patients with peanut and soybean allergies and seven non-atopic controls were examined. The majority of epitopes in Ara h 2 identified by microarray are consistent with those identified previously. Several regions in the 2S albumins are weakly recognized by individual sera from different patients. A comparison of allergenic epitopes on peanut and soybean proteins suggests that loop-helix type secondary structures and some amino acids with a large side chain including lone electron pair, such as arginine, glutamine, and tyrosine, makes the peptides highly recognizable by the immune system. By utilizing the peptide microarray assay, we mapped IgE epitopes of Ara h 2 and two soybean 2S albumins. The use of peptide microarray mapping and analysis of the epitope characteristics may provide critical information to access the allergenicity of food proteins.

Reactivity change of IgE to buckwheat protein treated with high-pressure and enzymatic hydrolysis.

BACKGROUND: Buckwheat is a popular food material in eastern Asian countries that can cause allergenic response. This study was conducted to evaluate the effects of hydrolysis with papain and high-pressure (HP) treatment of buckwheat protein (BWP) on reactivity of immunoglobulin E (IgE) and its secondary structure. RESULTS: Reactivity of IgE was examined by enzyme-linked immunosorbent assay (ELISA) with serum samples from 16 patients allergic to buckwheat. Reactivity of IgE to hydrolysate of BWP with papain showed a maximum decrease of 79.8%. After HP treatment at 600 MPa for 1 min, reactivity of IgE to BWP decreased by up to 55.1%. When extracted, BWP was hydrolyzed with papain overnight following HP treatment at 600 MPa which the reactivity of IgE decreased significantly by up to 87.1%. Significant changes in secondary structure of BWP were observed by circular dichroism (CD) analysis after hydrolysis with papain following HP treatment. CONCLUSION: Reduction of reactivity of IgE showed a correlation with changes in secondary structure of BWP, which may cause changes in conformational epitopes. This suggests the possibility of decreasing the reactivity of IgE to BWP using combined physical and enzymatic treatments.

Special consideration is required for the component-resolved diagnosis of egg allergy in infants.

BACKGROUND: There are few reports regarding differences in reactivity to the major egg allergens according to children's age, although component-resolved diagnosis is gradually being used. OBJECTIVE: To investigate differences in reactivity to major egg allergens among various age groups of children with egg allergy. METHODS: Twenty-seven patients diagnosed with egg allergy were included. Egg allergy was defined as a convincing history of reproducible symptoms within 2 hours of egg consumption and an egg white-specific IgE level of at least 0.35 kUA/L. Patients were divided into 3 age groups: younger than 12 months (group A, 7 subjects), 12 to 23 months (group B, 8 subjects), and at least 24 months (group C, 12 subjects). Immunoblotting and enzyme-linked immunosorbent assay investigated IgE reactivity toward ovalbumin, ovomucoid, and ovotransferrin in eggs. RESULTS: Immunoblotting analysis showed that all patients in group A reacted to ovalbumin, whereas reactions to other proteins were not detected. All patients in group B displayed a reaction to ovalbumin and ovomucoid. IgE binding to ovotransferrin was shown in 3 patients in group B. All patients in group C displayed reactivity to ovalbumin, 5 patients showed a reaction to ovomucoid, and 8 patients displayed a reaction to ovotransferrin. As a patient's age increased, specific IgE binding to ovalbumin and ovotransferrin increased (P = .011 and .004). CONCLUSION: IgE reactivity to egg allergens differs according to children's ages.

Infrared camera-proven water-damaged homes are associated with the severity of atopic dermatitis in children.

BACKGROUND: Moisture problems in dwellings have been linked to respiratory symptoms, but little is known about their association with symptoms of atopic dermatitis (AD). Moreover, the questionnaire-based survey or visual inspection for water damage does not use a standardized approach for assessing dampness. OBJECTIVE: To determine water damage in the houses of children with AD by assessing variations in surface temperature with an infrared camera, an interview-led questionnaire, and evaluation of the relation between the presence of water damage and the severity of AD. METHODS: Fifty-two homes of patients with AD were visited, and air samples were obtained from the living room and a child's bedroom. Water damage was determined by thermal assessments using an infrared camera and by the presence of visible mold or water stains. The effect of water damage on the severity of AD was analyzed by comparing the presence or absence of water damage and other aggravating factors between water-damaged and undamaged homes. RESULTS: Water damage was observed in 31 of 52 homes (59.6%), and the concentrations of airborne mold were significantly higher in water-damaged homes than in undamaged homes (P = .0013). However, there was no difference in airborne mold levels between homes with and those without visible mold or water stains. Logistic regression analyses showed that water-damaged homes were significantly related to moderate to severe AD (adjusted odds ratio 14.52, 95% confidence interval 1.75-121.13, P = .0025). CONCLUSION: Water-damaged homes affect the severity of AD in children. Infrared camera-driven assessment is a promising tool for determining moisture problems in buildings.