Preparng the right data diet for training neural networks

Neural networks are a good way to interrelate nonlinear variables in a robust manner. The simplex method for optimization is not nearly as effectual, and neither are the various statistical methods for classifying and associating data and predicting results. The reason is that neural networks are put through a training phase, during which they can automatically fine-tune themselves as often as proves necessary to get the desired performance. Of course, the old adage “garbage in...garbage out” applies as much to neural networks as it does to all other data-processing applications. If the training data set (the collection of input data and its associated correct output data) is not thoughtfully chosen, the resulting network is unlikely to hold up well in an industrial environment. So it is hardly surprising that massaging the set of training data consumes some 80 percent of the engineering time spent getting a real-world neural network up and running-that is, getting it to converge under a broad enough range of conditions to be deployed with confidence in a production situation. If that data preparation is done systematically, much time can be saved and a more robust end-product can be obtained. A nine-step process is given that experience (the author's) has shown can enhance the probability of obtaining a learning convergence robust enough for industrial use     

The future of Medicare. Their plans, your practice, Medicare reform, and how it will affect the practice of dermatology

OBJECTIVE: To investigate the relationship between adolescent risk behaviours, taking into account their influence upon one another. DESIGN: Cross-sectional survey utilising a self-completed questionnaire; stepwise logistic regression analyses were carried out, stratified for gender. SETTING: High schools in the Cape Peninsula, South Africa SUBJECTS: 7,340 students from 16 schools in the three major education departments. OUTCOME MEASURES: The outcome variables for each regression model were: cigarette smoking, cannabis smoking, alcohol bingeing, sexual intercourse, knife-carrying at school, walking home at night from beyond the neighbourhood, attempting suicide, and failure to use a seat belt. For each model, 26 risk behaviours served as independent variables. RESULTS: For each model, between 3 and 9 variables qualified for inclusion for each gender. There was a substantial association between many forms of substance abuse. In the previous 12 months, suicidal thoughts or statements of suicidal intent were predictors of a suicide attempt. Several variables involving injury were predictors of exposure to danger in getting home at night, and this was a predictor of substance abuse. Cannabis smoking, alcohol bingeing and exposure to danger in getting home at night were predictors of and were predicted by having had sexual intercourse. CONCLUSION: There are significant relationships between many adolescent risk behaviours, even when the influence of other risk behaviours is taken into account. The probability of adverse sequelae of risk behaviours, such as exposure to danger in getting home at night and sexual intercourse, is increased by the presence of selected other risk behaviours.     

The temporal coordination of early infant communication.

The ability to coordinate expressive behaviors is crucial to the development of social and emotional communication. Coordination involves systematic sequencing of behaviors from two different modalities that have some temporal overlap. A bootstrapping procedure was used to determine whether preverbal 3- and 6-month-old infants sequence vocalizations, gazes at their mothers' faces, and facial expressions into pairs of coordinated patterns nonrandomly. Smiles and frowns were highly coordinated with vocalizations. Smiles were also coordinated with gazes at mothers' faces, which became stronger with age. Vocalizations were not coordinated with gazes at mothers' faces. These findings illustrate the manner in which infants temporally coordinate communicative actions and provide new evidence that facial expressions (particularly smiles) are central to early infant communications. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Therapist experience, general clinical ability, and treatment outcome in schizophrenia.

Knowledge of the effect of therapist training and experience on the outcome of treatment of schizophrenic patients is scanty. The present article presents data systematically collected in the course of a controlled comparison of the effects of 5 different treatment methods in schizophrenia (individual psychotherapy, ataractic drug treatment, individual psychotherapy and drug treatment, ECT, and "milieu" care) involving 228 1st-admission schizophrenics without significant prior treatment and 38 psychiatric residents or recently graduated psychiatrists. Among the 23 outcome variables studied (including the Menninger Health-Sickness Scale, the Camarillo Dynamic Assessment Scale, the MMPI, the Communication subscale of the MACC Behavioral Adjustment Scale, and the Clyde Mood Scale), there was not a single instance in which the effect of therapist experience and general clinical ability was significantly related to outcome. There appeared to be, however, differences among therapists' results that were not related to experience and general clinical ability, particularly in relation to the length of time that they kept their patients in hospital. Drug treatment tended to override but perhaps not entirely eliminate these effects. (36 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Odor discrimination deficits in schizophrenia: association with eye movement dysfunction

Odor discrimination deficits were found in 80% of 20 schizophrenia patients and in none of the 20 age- and sex-matched comparison subjects. Olfactory discrimination was reliably measured in the patients. Twelve patients in this study also had smooth pursuit eye movement (SPEM) qualitatively recorded. The olfactory discrimination scores were highly correlated to SPEM but not to other clinical measures. This correlation suggests a shared neurobiology, possibly involving working memory.     

Hemorrhagic pleural effusion and pleural thickening as a complication of chronic lymphocytic leukemia

In the Upper Rhine Valley (Germany), Bacillus thuringiensis var. israelensis has been widely used against floodwater mosquitoes over an area of approximately 500 km2 for more than 10 years. The susceptibility of larvae of Aedes vexans field populations in 3 untreated (Lake Constance) and 3 treated areas (Upper Rhine Valley) was assessed by means of bioassays with B.t.i. (Bactimos WP, 6,000 AAU/mg), following WHO guidelines. Log-probit analyses and statistical evaluations of the data showed that the LC50 values as well as slopes of bioassays of the larvae deriving from the different areas showed no significant differences. Two populations in the treated area were even more susceptible than populations from the untreated areas. These results have been confirmed by resistance ratios, which were less than one in all tests carried out.     

不同纤维掺量下聚乙烯醇纤维/水泥复合材料徐变性能试验

通过对5组试件进行试验,研究了不同纤维掺量下聚乙烯醇纤维/水泥复合材料（PVA/ECC）的受压徐变性能。结果表明:PVA纤维的加入,降低了基材的密实度及弹性模量,使试件的徐变增加;在掺杂PVA纤维（体积分数为0~2vol%）的PVA/ECC试件中,PVA纤维掺量较高和掺量较低的试件徐变均较大,PVA纤维掺量适中的试件徐变较小;各组试件的徐变速率均具有前期快后期慢的特性,7天内发生的徐变可达总徐变的40%左右,60天内约为总徐变的80%,持荷约60天后,徐变逐渐趋于收敛。最后通过对试验值进行回归分析,提出了适用于PVA/ECC材料的徐变度数学计算模型及徐变系数预测模型。      

单相PWM整流器电压环的优化设计与仿真

为了改进单相电压型脉宽调制（PWM）整流器电压环的动态设计,提出了1种双零点典型Ⅱ型系统设计法,用于该整流器系统电压外环的调节器结构与参数优化设计,以取代目前广泛采用的典型Ⅱ型系统设计法。然后对按照这2种不同方法设计的系统进行了Simulink建模和仿真。仿真结果表明,采用提出的新方法设计的整流器系统时域性能指标更好。     

Live imaging of fluorescent proteins in chordate embryos: from ascidians to mice

Although we have advanced in our understanding of the molecular mechanisms intrinsic to the morphogenesis of chordate embryos, the question of how individual developmental events are integrated to generate the final morphological form is still unresolved. Microscopic observation is a pivotal tool in developmental biology, both for determining the normal course of events and for contrasting this with the results of experimental and pathological perturbations. Since embryonic development takes place in three dimensions over time, to fully understand the events required to build an embryo we must investigate embryo morphogenesis in multiple dimensions in situ. Recent advances in the isolation of naturally fluorescent proteins, and the refinement of techniques for in vivo microscopy offer unprecedented opportunities to study the cellular and molecular events within living, intact embryos using optical imaging. These technologies allow direct visual access to complex events as they happen in their native environment, and thus provide greater insights into cell behaviors operating during embryonic development. Since most fluorescent protein probes and modes of data acquisition are common across species, we have chosen the mouse and the ascidian, two model organisms at opposite ends of the chordate clade, to review the use of some of the current genetically-encoded fluorescent proteins and their visualization in vivo in living embryos for the generation of high-resolution imaging data.