Hybrid feature selection by combining filters and wrappers

Feature selection aims at finding the most relevant features of a problem domain. It is very helpful in improving computational speed and prediction accuracy. However, identification of useful features from hundreds or even thousands of related features is a nontrivial task. In this paper, we introduce a hybrid feature selection method which combines two feature selection methods – the filters and the wrappers. Candidate features are first selected from the original feature set via computationally-efficient filters. The candidate feature set is further refined by more accurate wrappers. This hybrid mechanism takes advantage of both the filters and the wrappers. The mechanism is examined by two bioinformatics problems, namely, protein disordered region prediction and gene selection in microarray cancer data. Experimental results show that equal or better prediction accuracy can be achieved with a smaller feature set. These feature subsets can be obtained in a reasonable time period.     

Adding enzymatically modified gelatin to enhance the rehydration abilities and mechanical properties of bacterial cellulose

Bacterial cellulose (BC) possesses excellent water holding capacity. However, increased crystallization causes a decrease in the rehydration ability of dried BC. Due to its excellent hydrophilic properties, we used gelatin and its enzymatically modified form (EMG) to prepare BC nano-composites in an attempt to enhance the rehydration abilities properties of BC. The polar peptide fraction was increased by extended hydrolysis of fish gelatin in Alcalase. Peptides smaller than 10 kDa were obtained by hydrolysis at 50 °C for 20 min. Protein contents of composites prepared by immersing BC in 5% gelatin (Gelatin/BC) or EMG (EMG/BC) solution were 81% and 92%, respectively, both of which then formed high gelatin/BC composites (HGBC). The protein content of EMG/BC was higher than that of Gelatin/BC, as compared with low gelatin/BC composites (LGBC) that were formed by immersion in a corresponding 0.5% solution. Among both HGBC and LGBC composites, EMG/BC exhibited the best rehydration abilities. Freeze-dried EMG/BC also exhibited the fastest rehydration rate and the best ability to restore wet-type composites. Composite microstructures revealed that BC was enveloped by gelatin when non-polar EMG (NPEMG) and polar EMG (PEMG) entered the BC network and adsorbed onto cellulose ribbons. The microstructure of EMG/BC contained both PEMG/BC and NPEMG/BC structures. Gelatin hydrolysates, penetrating BC networks and forming stable composites, improved the rehydration ability of dried BC. The polar functional groups of gelatin and its hydrolysates represent the key factors contributing to the hydrophilic nature of composites.     

Low molecular weight chitosan prepared with the aid of cellulase,lysozyme and chitinase: Characterisation and antibacterial activity

A continuous set of low molecular weight chitosan (LMWC) products was successfully made for this study by coordinating three enzymes (chitinase, lysozyme and cellulase) and two different deacetylated chitosan substrates (80% and 92%). It was observed that the intrinsic viscosity molecular weight (M_V) and SEC-HPLC-determined M_W distribution of LMWC were directed by both the used enzyme and the degree of chitosan substrate acetylation. LMWC prepared using lysozyme and 92%-deacetylated chitosan had larger M_W and, therefore, possessed higher antibacterial activity, as compared to other combinations. LMWC enzyme-directed properties suggest chitinase is more predictable and flexible to produce the specified M_V of LMWC. LMWC’s solubility and antibacterial activity, determined as minimum inhibitory concentration (MIC), against Escherichia coli exhibited a negative linear relationship with log M_V. E. coli strains showed much higher susceptibility to LMWCs than Staphylococcus aureus strains did. Both species also showed intra-species sensitivity diversity toward LMWC.     

Incremental backpropagation learning networks

How to learn new knowledge without forgetting old knowledge is a key issue in designing an incremental-learning neural network. In this paper, we present a new incremental learning method for pattern recognition, called the "incremental backpropagation learning network", which employs bounded weight modification and structural adaptation learning rules and applies initial knowledge to constrain the learning process. The viability of this approach is demonstrated for classification problems including the iris and the promoter domains.     

Real-time 3D human objects rendering based on multiple camera details

Multimedia Tools and Applications - 3D model construction techniques using RGB-D information have been gaining a great attention of the researchers around the world in recent decades. The RGB-D...     

Influence of alanine uptake on Staphylococcus aureus surface charge and its susceptibility to two cationic antibacterial agents,nisin and low molecular weight chitosan

Low molecular weight chitosan (LMWC) and nisin, recognized as cationic antibacterial agents (CAAs), inhibit bacterial growth by interacting with the anionically charged cell wall. In this study, alanine uptake significantly reduced the anionic cell surface charge, as determined by the zeta potential (ZP) measurements, of Staphylococcus aureus, resulting from the incorporation of d-alanine into the cell wall. Minimum inhibitory concentration (MIC) tests and growth inhibition curves revealed that LMWC and nisin possessed inverse antibacterial activity against three strains of S. aureus, depending on the strains’ net charge. A twofold reduction in the MIC value of nisin was obtained against S. aureus, inoculated in a 1.0% d- or l-alanine-augmented trypticase soy broth medium. A flocculation test demonstrated that neutralizing the anionic surface charge using d-alanine reduced the adsorption of S. aureus onto LMWC. Furthermore, the reduced surface net charge could enhance the colonization capacity of S. aureus on glass.     

免疫营养在COVID-19综合治疗中的辅助作用与应用前景

当前,由新型冠状病毒SARS-CoV-2引起的新型冠状病毒肺炎(COVID-19)大流行已在全球蔓延。除了隔离之外,迫切需要采取能够降低感染和死亡风险的公共卫生措施。营养不良和导致患者免疫功能低下的各种非传染性基础疾病(如肥胖症、糖尿病、心血管疾病等)是重症COVID-19的重要危险因素。应用免疫营养手段,补充具有药理学作用的特殊营养素,以特定方式刺激免疫细胞,增强免疫应答功能,维持适度的免疫反应,调整细胞因子的产生和释放,减轻有害或过度炎症反应,对COVID-19患者预后具有积极意义。本文就维生素、矿物元素、多不饱和脂肪酸、膳食纤维等免疫营养素在降低呼吸道感染风险中的作用,以及免疫营养在COVID-19中应用现状做一综述,以期为免疫营养在COVID-19综合治疗中的辅助作用提供可靠参考依据。     

Isolation and characterization of acid and pepsin-solubilized collagens from the skin of balloon fish (Diodon holocanthus)

Acid-solubilized collagen (ASC) and pepsin-solubilized collagen (PSC) were successfully extracted from the skin of balloon fish (Diodon holocanthus) with yields of 4% and 19.5% respectively (based on dry weight). According to the electrophoretic patterns, both the ASC and PSC consisted of two different α chains (α1 and α2), were characterized to be type I, and contained imino acid of 179 and 175 residues/1000 residues, respectively. The PSC had a lower content of high-molecular weight cross-links than the ASC. The ultraviolet (UV) absorption spectrum of collagens showed that the distinct absorption was between 210 and 230 nm. A maximum solubility in 0.5 M acetic acid was observed at pH 1–5, and a sharp decrease in solubility above 4% (w/v) in both the ASC and PSC was observed in the presence of NaCl. The denaturation temperatures (T_d) of the ASC and PSC measured by viscometry were 29.01 °C and 30.01 °C, respectively. The maximum temperatures (T_max) of the ASC and PSC were 29.64 °C and 30.30 °C, respectively.     

压滤机主要受力构件强度计算的探讨

就压滤机主要受力构件的强度计算,做了几点探讨,主要体现在许用挠度的取值和深梁的计算上,旨在为基层提供较实用的设计依据。     

基于嵌入式的水产养殖在线监测系统

针对传统的水产养殖向工厂化、集约化发展过程中监测手段薄弱等问题,提出一种水产养殖在线监测系统的设计方案。该系统基于嵌入式技术、结合GSM/GPRS通信与射频无线传输技术,主要实现对水产品生长的水环境因素（温度、PH 值和溶解氧含量）进行实时监测。智能化的设计使监测人员根据异常数据及时采取有效措施。实验证明：系统结构简单,功能易扩展,实测数据误差小,适用于工厂化的水产养殖领域。