Search

CTR prediction is an important research direction in the field of recommendation system. In recent years, DeepFM has been concerned because it can learn the low- and high-order feature interactions end-to-end, and it has become one of the best models for CTR prediction. However, due to the inadequate feature interactions in the FM component of DeepFM model, the prediction accuracy is reduced. Compared with DeepFFM, although the FFM component of DeepFFM model attributes the same feature to the same filed to alleviates the inadequate feature interactions in FM model, its binomial parameter number is far greater than DeepFM, which increases the pressure of recommendation system. In this paper, we propose a new model called DeepFaFM. It provides a scalable coding matrix with adjustable factors, which can dynamically balance the operation pressure according to the actual production situation. Compared with DeepFM, which is one of the latest classification models, it has the advantages of DeepFM, improves the inadequate feature interactions in the operation of DeepFM, and improves the accuracy of the model. Experiments show that DeepFaFM is better than DeepFM in accuracy, and is equivalent to DeepFFM but its operation efficiency is significantly higher than DeepFFM. In particular, DeepFaFM has a scalable variable, which can dynamically balance the accuracy and calculation pressure according to the actual production situation, improving the flexibility of the model.

BACKGROUND/AIM Recent studies have shown that low haemoglobin (Hb) levels promote the progression of chronic kidney disease. This study assessed the relationship between Hb level and type 1 diabetic nephropathy (DN) in Han patients in Anhui, China. METHODS There was a total of 236 patients diagnosed with type 1 diabetes mellitus (T1DM) seen between January 2014 and December 2016 in our centre. Haemoglobin levels in patients with DN were compared with those without DN. The relationship between Hb level and the urinary albumin-creatinine ratio (ACR) was examined by Spearman's correlational analysis and multiple stepwise regression analysis. The binary logistic multivariate regression analysis was performed to analyse the correlated factors for type 1 DN, calculate the odds ratio (OR) and 95% confidence interval (CI). The predicting value of Hb level for DN was evaluated by area under receiver operation characteristic curve (AUROC) for discrimination and Hosmer-Lemeshow goodness-of-fit test for calibration. RESULTS The average Hb levels in the DN group (116.1 ± 20.8 g/L) were significantly lower than the non-DN group (131.9 ± 14.4 g/L), P < 0.001. Hb levels were independently correlated with the urinary ACR in multiple stepwise regression analysis. The logistic multivariate regression analysis showed that the Hb level (OR: 0.936, 95% CI: 0.910-0.963, P < 0.001) was inversely correlated with DN in patients with T1DM. In sub-analysis, low Hb level (Hb < 120 g/L in female, Hb < 130 g/L in male) was still negatively associated with DN in patients with T1DM. The AUROC was 0.721 (95% CI: 0.655-0.787) in assessing the discrimination of the Hb level for DN. The value of P was 0.593 in Hosmer-Lemeshow goodness-of-fit test. CONCLUSIONS In patients with T1DM, the Hb level is inversely correlated with urinary ACR and DN.

Malnutrition is the main determinant of mortality and morbidity in maintenance hemodialysis patients. In many countries except for China, it has been reported that short daily hemodialysis (SDHD) could improve nutritional status. We will report here the nutritional results obtained in the SDHD therapy period compared with conventional hemodialysis (cHD) therapy period in Chinese patients. This study compared 29 SDHD patients (SDHD group), each patient served as his own control, with 30 cHD patients (cHD group) serving as the parallel controls. The hematologic parameters, anthropometric measurements, modified quantitative subjective global assessment (MQSGA) score, weekly standard Kt/V (std Kt/V) and average daily intake of protein were measured at baseline (SDHD0 or cHD0 period), at 3 months (SDHD1 or cHD1 period) and at 6 months (SDHD2 or cHD2 period). The average daily intake of protein, dry weight, body mass index, mid-arm circumference, mid-arm muscle circumference, serum albumin, prealbumin, cholesterol, hemoglobin, weekly std Kt/V values at SDHD2 were higher than the corresponding values at SDHD0 (p

In recent years, the chimeric antigen receptor modified T cells (Chimeric antigen receptor T cells, CAR-T) immunotherapy has developed rapidly, which has been considered the most promising therapy. Efforts to enhance the efficacy of CAR-based anti-tumor therapy have been made, such as the improvement of structures of CAR-T cells, including the development of extracellular antigen recognition receptors, intracellular co-stimulatory molecules and the combination application of CARs and synthetic small molecules. In addition, effects on the function of the CAR-T cells that the space distance between the antigen binding domains and tumor targets and the length of the spacer domains have are also being investigated. Given the fast-moving nature of this field, it is necessary to make a summary of the development of CAR-T cells. In this review, we mainly focus on the present design strategies of CAR-T cells with the hope that they can provide insights to increase the anti-tumor efficacy and safety.

Objective: This study assessed the relationship between serum albumin (ALB) at start of peritoneal dialysis (PD) and long-term outcomes of continuous ambulatory PD (CAPD) in Anhui Han patients. Methods: A total of 149 Anhui Han CAPD patients were enrolled in this study and followed up for 3 years. They were initially diagnosed with the end-stage renal disease and underwent surgical PD catheter placement from January 2009 to December 2013. According to serum ALB at start of PD, the patients were divided into two groups: low ALB group (ALB < 35 g/L) and high ALB group (ALB ≥35 g/L). Demographic, hematologic, biochemical, and dialysis-related data were collected. Kaplan-Meier survival analysis and log-rank test were conducted to compare patient mortality, cardiovascular mortality and technique failure between the low ALB group and the high ALB group. Cox regression analysis was performed to analyze the risk factors, calculate the hazard ratio (HR), adjusted HR (AHR) and 95% confidence interval (CI). Results: The low ALB group showed a greater number of diabetes mellitus compared with the high ALB group. Patient mortality, cardiovascular mortality, and technique failure in the high ALB group were significantly lower than those in the low ALB group. In Cox regression analysis, serum ALB < 35 g/L was an independent predictor of patient mortality (AHR 3.043, 95% CI 1.085–8.536, p = 0.034), cardiovascular mortality (AHR 11.587, 95% CI 1.466–91.574, p = 0.020), and technique failure (AHR 3.148, 95% CI 1.603–6.182, p = 0.001) in CAPD patients after adjustment for sex, age, estimated glomerular filtration rate, primary renal disease, diabetes mellitus, and cardiovascular disease. Conclusions: In Anhui Han patients on CAPD, the levels of serum ALB at start of PD are inversely correlated with patient mortality, cardiovascular mortality, and technique failure, and the long-term outcomes of patients with hypoalbuminemia at start of PD are poor. To improve the long-term outcomes of Anhui Han CAPD patients, patients with hypoalbuminemia at start of PD should be closely monitored.

Recent studies have shown that low haemoglobin (Hb) levels promote the progression of chronic kidney disease. This study assessed the relationship between Hb level and type 1 diabetic nephropathy (DN) in Han patients in Anhui, China.There was a total of 236 patients diagnosed with type 1 diabetes mellitus (T1DM) seen between January 2014 and December 2016 in our centre. Haemoglobin levels in patients with DN were compared with those without DN. The relationship between Hb level and the urinary albumin-creatinine ratio (ACR) was examined by Spearman's correlational analysis and multiple stepwise regression analysis. The binary logistic multivariate regression analysis was performed to analyse the correlated factors for type 1 DN, calculate the odds ratio (OR) and 95% confidence interval (CI). The predicting value of Hb level for DN was evaluated by area under receiver operation characteristic curve (AUROC) for discrimination and Hosmer-Lemeshow goodness-of-fit test for calibration.The average Hb levels in the DN group (116.1 ± 20.8 g/L) were significantly lower than the non-DN group (131.9 ± 14.4 g/L), P0.001. Hb levels were independently correlated with the urinary ACR in multiple stepwise regression analysis. The logistic multivariate regression analysis showed that the Hb level (OR: 0.936, 95% CI: 0.910-0.963, P0.001) was inversely correlated with DN in patients with T1DM. In sub-analysis, low Hb level (Hb120 g/L in female, Hb130 g/L in male) was still negatively associated with DN in patients with T1DM. The AUROC was 0.721 (95% CI: 0.655-0.787) in assessing the discrimination of the Hb level for DN. The value of P was 0.593 in Hosmer-Lemeshow goodness-of-fit test.In patients with T1DM, the Hb level is inversely correlated with urinary ACR and DN.