GARCHNet: Value-at-Risk Forecasting with GARCH Models Based on Neural Networks.

This paper proposes a new GARCH specification that adapts the architecture of a long-term short memory neural network (LSTM). It is shown that classical GARCH models generally give good results in financial modeling, where high volatility can be observed. In particular, their high value is often praised in Value-at-Risk. However, the lack of nonlinear structure in most approaches means that conditional variance is not adequately represented in the model. On the contrary, the recent rapid development of deep learning methods is able to describe any nonlinear relationship in a clear way. We propose GARCHNet, a nonlinear approach to conditional variance that combines LSTM neural networks with maximum likelihood estimators in GARCH. The variance distributions considered in the paper are normal, t and skewed t, but the approach allows extension to other distributions. To evaluate our model, we conducted an empirical study on the logarithmic returns of the WIG 20 (Warsaw Stock Exchange Index), S&P 500 (Standard & Poor's 500) and FTSE 100 (Financial Times Stock Exchange) indices over four different time periods from 2005 to 2021 with different levels of observed volatility. Our results confirm the validity of the solution, but we provide some directions for its further development.

The Advantage of Case-Tailored Information Metrics for the Development of Predictive Models, Calculated Profit in Credit Scoring.

This paper compares model development strategies based on different performance metrics. The study was conducted in the area of credit risk modeling with the usage of diverse metrics, including general-purpose Area Under the ROC curve (AUC), problem-dedicated Expected Maximum Profit (EMP) and the novel case-tailored Calculated Profit (CP). The metrics were used to optimize competitive credit risk scoring models based on two predictive algorithms that are widely used in the financial industry: Logistic Regression and extreme gradient boosting machine (XGBoost). A dataset provided by the American Fannie Mae agency was utilized to conduct the study. In addition to the baseline study, the paper also includes a stability analysis. In each case examined the proposed CP metric that allowed us to achieve the most profitable loan portfolio.

A Probability-Based Models Ranking Approach: An Alternative Method of Machine-Learning Model Performance Assessment.

Performance measures are crucial in selecting the best machine learning model for a given problem. Estimating classical model performance measures by subsampling methods like bagging or cross-validation has several weaknesses. The most important ones are the inability to test the significance of the difference, and the lack of interpretability. Recently proposed Elo-based Predictive Power (EPP)-a meta-measure of machine learning model performance, is an attempt to address these weaknesses. However, the EPP is based on wrong assumptions, so its estimates may not be correct. This paper introduces the Probability-based Ranking Model Approach (PMRA), which is a modified EPP approach with a correction that makes its estimates more reliable. PMRA is based on the calculation of the probability that one model achieves a better result than another one, using the Mixed Effects Logistic Regression model. The empirical analysis was carried out on a real mortgage credits dataset. The analysis included a comparison of how the PMRA and state-of-the-art k-fold cross-validation ranked the 49 machine learning models, an example application of a novel method in hyperparameters tuning problem, and a comparison of PMRA and EPP indications. PMRA gives the opportunity to compare a newly developed algorithm to state-of-the-art algorithms based on statistical criteria. It is the solution to select the best hyperparameters configuration and to formulate criteria for the continuation of the hyperparameters space search.

Multiclass Image Classification Using GANs and CNN Based on Holes Drilled in Laminated Chipboard.

The multiclass prediction approach to the problem of recognizing the state of the drill by classifying images of drilled holes into three classes is presented. Expert judgement was made on the basis of the quality of the hole, by dividing the collected photographs into the classes: "very fine," "acceptable," and "unacceptable." The aim of the research was to create a model capable of identifying different levels of quality of the holes, where the reduced quality would serve as a warning that the drill is about to wear down. This could reduce the damage caused by a blunt tool. To perform this task, real-world data were gathered, normalized, and scaled down, and additional instances were created with the use of data-augmentation techniques, a self-developed transformation, and with general adversarial networks. This approach also allowed us to achieve a slight rebalance of the dataset, by creating higher numbers of images belonging to the less-represented classes. The datasets generated were then fed into a series of convolutional neural networks, with different numbers of convolution layers used, modelled to carry out the multiclass prediction. The performance of the so-designed model was compared to predictions generated by Microsoft's Custom Vision service, trained on the same data, which was treated as the benchmark. Several trained models obtained by adjusting the structure and hyperparameters of the model were able to provide better recognition of less-represented classes than the benchmark.

Prevalence of hypogammaglobulinemia and its management with subcutaneous immunoglobulin supplementation in patients after allogeneic hematopoietic stem cell transplantation-a single-center analysis.

Secondary immunodeficiencies are frequently observed after allo-HSCT. The efficacy of subcutaneous IgG preparations in this population is unknown. A retrospective single-institution study involved 126 adult patients transplanted in 2012-2019 for hematological malignancies. Patients were tested every 2-3 weeks for plasma IgG concentration during the 1st year after transplantation and supplemented with facilitated subcutaneous immunoglobulin when they either had IgG concentration < 500 mg/dl or between 500 and 700 mg/dl and recurrent infection. The IgG concentration < 500 mg/dL was diagnosed in 41 patients, while 500-700 mg/dL in 25 and altogether 53 patients received IgG supplementation. The median number of IgG administrations was 2. The median time to the first IgG administration after allo-HSCT was 4.1 months, while to the next administration (if more than one was required) 53 days (prophylactic group) and 32 days (group with infections). We did not observe any significant toxicity. Two situations were associated with increased probability of meeting criteria for IgG supplementation: diagnosis of either acute lymphoblastic leukemia (ALL) or chronic lymphocytic leukemia (CLL) (83.8% versus 39.3% for other diagnosis, p = 0.000) and the systemic use of corticosteroids (64.2% versus 31.5% for patients without systemic corticosteroids, p = 0.005). Over 40% of the adult recipients may require at least incidental immunoglobulin supplementation during the first year after allo-HSCT. Low IgG concentrations are associated with inferior outcomes. The subcutaneous route of IgG administration appeared to be safe and may allow for long persistence.

Relationship between age and in-hospital mortality during 15,345,025 non-surgical hospitalizations.

INTRODUCTION: Mortality, whether in or out of hospital, increases with age. However, studies evaluating in-hospital mortality in large populations did not distinguish between surgical and non-surgical causes of death, either in young or in elderly patients. The aim of the study was to assess in-hospital non-surgical mortality in a large group of patients, with a special focus on the elderly. MATERIAL AND METHODS: Data from the database of the Polish National Health Fund (NHF) regarding hospitalizations of adult (≥ 18 years) patients not related to surgical procedures in the years 2009-2013 were used to assess in-hospital mortality. RESULTS: 15,345,025 hospitalizations were assessed. The mean in-hospital non-surgery-related mortality rate was 3.96 ±0.17%, and increased from 3.79% to 4.2% between 2009 and 2013. The mean odds ratio for in-hospital death increased with the age of patients, reaching a 229-fold higher rate in the ≥ 95 years age group as compared to the 18-24 age group. The highest mean mortality was associated with respiratory diseases (6.91 ±0.20%), followed by heart and vascular diseases, nervous system diseases, as well as combined gastrointestinal tract, liver, biliary tract, pancreas and spleen diseases (5.65 ±0.27%, 5.46 ±0.05% and 4.01 ±0.13%, respectively). CONCLUSIONS: The in-hospital non-surgery-related mortality rate was approximately 4%. It significantly increased with age and, regardless of age, was highest in patients suffering from respiratory diseases.

Predictors of In-Hospital Mortality in Surgical Wards: A Multivariable Retrospective Cohort Analysis of 2,800,069 Hospitalizations.

BACKGROUND: Identifying prognostic factors that are predictive of in-hospital mortality for patients in surgical units may help in identifying high-risk patients and developing an approach to reduce mortality. This study analyzed mortality predictors based on outcomes obtained from a national database of adult patients. MATERIALS AND METHODS: This retrospective study design collected data obtained from the National Health Fund in Poland comprised of 2,800,069 hospitalizations of adult patients in surgical wards during one calendar year. Predictors of mortality which were analyzed included: the patient's gender and age, diagnosis-related group category assigned to the hospitalization, length of the hospitalization, hospital type, admission type, and day of admission. RESULTS: The overall mortality rate was 0.8%, and the highest rate was seen in trauma admissions (24.5%). There was an exponential growth in mortality with respect to the patient's age, and male gender was associated with a higher risk of death. Compared to elective admissions, the mortality was 6.9-fold and 15.69-fold greater for urgent and emergency admissions (p < 0.0001), respectively. Weekend or bank holiday admissions were associated with a higher risk of death than working day admissions. The "weekend" effect appears to begin on Friday. The highest mortality was observed in less than 1 day emergency cases and with a hospital stay longer than 61 days in any type of admission. CONCLUSION: Age, male gender, emergency admission, and admission on the weekend or a bank holiday are factors associated with greater mortality in surgical units.

Reduced Albuminuria and Potassemia Indicate Early Renal Repair Processes after Resynchronization Therapy in Cardiorenal Syndrome Type 2.

BACKGROUND: Patients with chronic cardiorenal syndrome type 2 (T2-CRS) who qualify for resynchronization therapy (CRT) are exposed perioperatively to potentially nephrotoxic factors including contrast agents and blood loss. METHODS: The objective of this prospective interventional study was to assess the effects of CRT on renal function in patients with T2-CRS within the first 48 hours following implantation. Initially, 76 patients (15% female; aged 69 ± 9.56 years) with heart failure (New York Heart Association classes II-IV), ejection fraction ≤ 35%, and QRS > 130 ms were included in the study. During CRT implantation, a nonionic contrast agent (72.2 ± 44.9 mL) was administered. Prior to and 48 hours following implantation, renal function was evaluated using the following serum biomarkers: creatinine (sCr), estimated glomerular filtration rate (using the Chronic Kidney Disease Epidemiology Collaboration equation [eGFRCKD-EPI]), and the electrolyte and urine biomarkers albumin (uAlb), albumin/creatinine ratio (UACR), and neutrophil gelatinase-associated lipocalin (uNGAL). RESULTS: Before CRT, patients classified as NYHA class III or IV had higher uNGAL levels in comparison to uNGAL levels after CRT (43.63 ± 60.02 versus 16.63 ± 18.19; p=0.041). After CRT implantation, uAlb, UACR, and potassium levels were reduced (p < 0.05), and uNGAL, sCr, and eGFRCKD-EPI were unchanged. The contrast medium volume did not correlate with the test biomarkers (p > 0.05). CONCLUSIONS: In patients with T2-CRS, uNGAL is a biomarker of kidney injury that correlates with the NYHA classes. A stable uNGAL value before and after CRT implantation confirms the lack of risk of contrast-induced nephropathy. Reduced albuminuria and blood potassium are biomarkers of improving T2-CRS in the early post-CRT period.

Predictors of in-hospital mortality in nonsurgical departments: a multivariable regression analysis of 2 855 029 hospitalizations.

INTRODUCTION: In­hospital mortality is a relevant outcome of hospital admissions. OBJECTIVES: This study aimed to identify predictors independently associated with in­hospital mortality in nonsurgical departments. PATIENTS AND METHODS: In 2014, the Polish National Health Fund database provided data on 2 855 029 hospitalizations of adults, which were not related with surgical procedures. Patients' age and sex, diagnosis­related group category assigned to the hospitalization, length of stay, types of hospital and admission, and day of the week and month of admission were analyzed as mortality predictors. RESULTS: The mean (SD) in­hospital mortality rate was 4.1% (0.01%). Odds ratios for in­hospital death increased with patients' age. The female sex was associated with lower odds ratios of death than the male sex. Among the diagnosis­related groups assigned to hospitalizations, the highest mortality was found in patients with vascular diseases (11.95%). Considering the length of stay, the lowest mortality occurred during 5- to 7­day (2.63%). Compared with teaching hospitals, the odds ratio of death was 1.31­fold higher for regional hospitals, 1.35­fold higher for private hospitals, and 1.48­fold higher for district and city hospitals; 92% of all in­hospital deaths occurred after urgent and emergency admissions. Hospital admissions at weekends or on other nonworking days (public holidays) were significant predictors of in­hospital mortality. Differences in mortality rates were found between particular months, but no seasonal relationship was established. CONCLUSIONS: Age, male sex, emergency admission, admission at the weekend or on another nonworking day (during public holidays), and hospitalization in a district, city, private, or regional hospital (compared with a university hospital) were factors associated with higher mortality in nonsurgical departments.

The role of adolescent acne treatment in formation of scars Among patients with persistent adult acne: evidence from an observational study.

Persistent adult acne is one of the most difficult types of acne to treat. It is a long-lasting disease with uncontrolled exacerbations that often result in scarring. The aim of this study was to analyze the influence of acne therapy used in adolescence on patients who later developed persistent adult acne. The use of oral antibiotics, isotretinoin, and topical retinoids in adolescence and their role in diminishing scar formation during adult acne was analyzed. This population-based study included 111 patients, 91 of whom had persistent adult acne. Results indicated that the use of isotretinoin or topical retinoids for adolescent acne decreased the risk for scar occurrence in adulthood.

Randomized comparison of two methods of the epidural space identification during regional labour analgesia.

OBJECTIVES: Conventional loss of resistance (LOR) technique for identifying the epidural space (EDS) predominantly depends on experience of the anaesthetist. A technique using automated syringe for EDS identification was invented as an alternative to the traditional method. The aim of the study was to compare the efficacy and risk for complications between automatic LOR syringe - Epimatic® (Vygon, Ecouen, France) and conventional LOR - Perifix® (B.Braun Melsungen AG, Melsungen, Germany) techniques for EDS identification. MATERIAL AND METHODS: A total of 170 patients were enrolled into the study and 153 cases were analysed. Number of at- tempts, time to EDS identification, ease of EDS identification, complication rate and patient procedure-related discomfort were evaluated and compared. RESULTS: No statistically significant differences were found in the number of needle insertion attempts (1.3 in both groups), time to EDS identification (31 sec. vs. 27 sec.), efficacy of epidural analgesia (100% in both groups), or complication rate between both groups. CONCLUSIONS: The automatic and the conventional LOR techniques are comparable in terms of efficacy and safety for the epidural space identification.

Is maintenance treatment in adult acne important? Benefits from maintenance therapy with adapalene, and low doses of alpha and beta hydroxy acids.

Background: Adult acne is a chronic disease with uncontrolled exacerbations, associated with a psychological burden of the patient and medical expenses. Aims: The aim of the study was to check the efficacy of maintenance therapy of adult acne. It is essential part of treatment as adult acne usually has a long-lasting and recurring course. Methods: In the study, the efficacy of maintenance therapy in patients with adult acne is evaluated. In this study, 100 patients (aged 25-39 years of age) with mild and moderate adult acne were enrolled. Results: The maintenance therapy (adapalene 0.1% three times a week and low doses of alpha and beta hydroxy acids) led to a significant decrease in the number of acne lesions (from 31.3 to 12.25; p < .001) and severity of seborrhea (p < .001). Conclusions: Maintenance therapy brings significant improvements in the reduction of non-inflammatory and inflammatory lesions in patients with mild and moderate adult acne.

Matrix metalloproteinases-2, -7 and tissue metalloproteinase inhibitor-1 expression in human endometrium.

INTRODUCTION: Endometrium undergoes regular, cyclic tissue remodeling mostly associated to the endocrine system status. It is well-known fact that steroid hormones are strongly responsible for changes in endometrium. The precise mechanism of their action is still under investigation. The aim of the study was to evaluate the expression of metalloproteinases 2 and 7 (MMP-2, -7) and tissue inhibitor of metalloproteinase 1 (TIMP-1) in human endometrium in relation to serum concentrations of estradiol and progesterone during different phases of menstrual cycle. MATERIAL AND METHODS: The study material consisted of 52 biopsy samples; 12 obtained in the proliferative phase, 11 in the secretory phase and 29 during menstruation. Expression of MMP-2, MMP-7 and TIMP-1 was assessed by immunohistochemistry. Serum concentrations of estradiol and progesterone at time of biopsy were evaluated by immunochemistry assay. Results of the study were statistically assessed by linear regression model. RESULTS: Increased serum concentration of estradiol was associated with increased MMP-2 expression in proliferative phase but decreased in secretory phase and during menstruation. No significant relationship was found between progesterone concentration and MMP-2 expression. Moreover, no difference in the expression of MMP-7 and TIMP-1 in the endometrium in relation to hormone levels and menstrual cycle phases were observed. CONCLUSIONS: The results of the study indicate that estradiol influence MMP-2 expression in the endometrium depends on the phase of menstrual cycle. Such relationships were not found for MMP-7 and TIMP-1 and further tests clarifying association between estradiol and MMPs are needed.

Expression of matrix metalloproteinase enzymes in endometrium of women with abnormal uterine bleeding.

OBJECTIVES: Abnormal uterine bleeding (AUB) is caused by derangement of physiological processes of tissue growth, shedding and regeneration. It is known that interplay between metalloproteinases (MMP's) and tissue inhibitors of metalloproteinases (TIMP's) may play a crucial role in its occurrence. AIM: To define if expression of proMMP-2, MMP-2 and TIMP-1 in endometrium of women with AUB is dependent on steroid sex hormone concentration and histopathological picture. MATERIALS AND METHODS: Endometrial scraps were taken from 21 women with AUB and 19 controls. Samples were evaluated in light microscopy by a certified pathologist. Activity of proMMP-2 and MMP-2 proteins levels were evaluated by gelatin zymography and TIMP-1 by reversed zymography. The results has been correlated with serum estradiol and progesterone concentrations in linear regression model. RESULTS: Expression: of proMMP-2 in endometrium of women with AUB is correlated with estradiol concentration and inversely correlated with progesterone levels. It was significantly higher in women with dysfunctional endometrium (p<0.001). Expression of MMP-2 was highest in women with endometrial polyps and longer bleeding (p<0.01), while expression of TIMP-1 was independent from hormone concentration. CONCLUSION: Lack of correlation between proMMP-2 and MMP-2 levels suggest different pathway of their activation in AUB. ProMMP-2 is up regulated by estradiol and down regulated by progesterone while MMP-2 levels increase with the length of bleeding.

Fetal Hypotrophy Is an Important Marker in Diagnosis of Preeclampsia in Pregnant Patients After Solid Organ Transplantation.

BACKGROUND The purpose of this study was to use a multidisciplinary approach to define the importance of fetal growth disturbances in pregnant patients after renal or liver transplantation in diagnosis and treatment of preeclampsia. MATERIAL AND METHODS We assessed 108 pregnancies in patients with renal or liver transplants. Statistical analysis included Pearson's chi-square test and Fisher's exact test. RESULTS In the renal transplant (RTR) group, preeclampsia was diagnosed in 40% according to ISSHP. In the liver transplant (LTR) group, ISSHP guidelines allow this diagnose in 14.6% of patients. Intrauterine fetal hypotrophy occurred in 53.3% of RTR patients with clinical symptoms of preeclampsia and in none of stabile patients. Premature delivery rate was 40% in patients with hypotrophy and only in 15.5% without. For LTR patients, hypotrophy was diagnosed in 16.4% patients with clinical symptoms of preeclampsia and in 12.7% of stabile patients. Premature delivery rate was 14.5% in patients with hypotrophy and in 14.5% without. CONCLUSIONS Fetal hypotrophy is strongly associated with premature delivery and risk of preeclampsia in pregnancies after renal transplantation. There is a need for including ultrasound findings in diagnostic criteria of preeclampsia. Fetal growth monitoring may help in prediction of premature delivery in these group.

Factors affecting the course and severity of adult acne. Observational cohort study.

OBJECTIVE: To identify factors improving symptoms and shortening duration of AA. MATERIAL AND METHODS: The observational cohort study was performed in 111 patients with AA (>25 y.o.) in 2015-2016. Clinical manifestation, previous treatments, environmental risk factors and features of juvenile acne affecting AA were assessed. RESULTS: The maximum severity of persistent acne was significantly lower after 25 years of age, as compared to adolescence (7.2 vs. 6.4; p = .0027). The number of acne therapies used in AA was twice as high as in juvenile acne (22 vs. 11). The severity of AA sufficient to leave scars was significantly lower than of juvenile acne (6.0 vs. 7.3; p = .0001) with 22% of patients developing scars only in adult life. Patients linked exacerbations to stress exposure (p = .09 and <.0001 for those reporting at least one stressor and all patients, respectively), finding lifestyle changes the most stressful (p = .046). Those using full-coverage foundations received significantly more acne treatments over lifetime (5.4 vs. 3.6; p = .0359) and for AA (4.4 vs. 2.8; p = .0043). Discontinuation of oral contraceptives or sensitive, erythema-prone skin also worsened the symptoms. CONCLUSION: Lifestyle change-related stress, sensitive skin, discontinuation of oral contraceptives and using full-coverage foundations increase severity of AA.

Mesenchymal Stem Cells from Human Amniotic Membrane and Umbilical Cord Can Diminish Immunological Response in an in vitro Allograft Model.

BACKGROUND/AIM: Mesenchymal stem cells (MSCs) are gaining rising interest in gynecology and obstetrics. MSCs immunomodulatory properties are suitable enough to reduce perinatal morbidity caused by inflammation in premature neonates. The aim of this study was to evaluate and compare the ability to inhibit allo-activated lymphocytes proliferation by MSCs derived from different sources: amniotic membrane (AM), umbilical cord (UC) and adipose tissue (AT). METHODS: MSCs were isolated from AM (n = 7) and UC (n = 6) and AT (n = 6) of healthy women. Cells were characterized by flow cytometry and differentiation assay. To evaluate the potential of fetal and adult MSCs to diminish immunological response, mixed lymphocytes reaction (MLR) was performed. RESULTS: Amnion and UC-derived cells displayed typical MSCs characteristics. Addition of MSCs to MLR significantly inhibited the proliferation of stimulated lymphocytes. The effect was observed regardless of the MSCs type used (p < 0.01 in all groups). Comparative analysis revealed no significant differences in this action between tested MSCs types. Additionally, no type of MSCs significantly stimulated allogeneic lymphocytes. CONCLUSION: The results prove the immunosuppressive capacities of fetal-derived MSCs in vitro. In the future, they may be potentially used to treat premature newborn as well as in immunomodulation in post-transplant therapy.

Comparison of the techniques for the identification of the epidural space using the loss-of-resistance technique or an automated syringe - results of a randomized double-blind study.

BACKGROUND: The conventional, loss of resistance technique for identification of the epidural space is highly dependent on the anaesthetist's personal experience and is susceptible to technical errors. Therefore, an alternative, automated technique was devised to overcome the drawbacks of the traditional method. The aim of the study was to compare the efficacy of epidural space identification and the complication rate between the two groups - the automatic syringe and conventional loss of resistance methods. METHODS: 47 patients scheduled for orthopaedic and gynaecology procedures under epidural anaesthesia were enrolled into the study. The number of attempts, ease of epidural space identification, complication rate and the patients' acceptance regarding the two techniques were evaluated. RESULTS: The majority of blocks were performed by trainee anaesthetists (91.5%). No statistical difference was found between the number of needle insertion attempts (1 vs. 2), the efficacy of epidural anaesthesia or the number of complications between the groups. The ease of epidural space identification, as assessed by an anaesthetist, was significantly better (P = 0.011) in the automated group (87.5% vs. 52.4%). A similar number of patients (92% vs. 94%) in both groups stated they would accept epidural anaesthesia in the future. CONCLUSION: The automated and loss of resistance methods of epidural space identification were proved to be equivalent in terms of efficacy and safety. Since the use of the automated technique may facilitate epidural space identification, it may be regarded as useful technique for anaesthetists inexperienced in epidural anaesthesia, or for trainees.