基于神经网络的多元智能

神经网络作为人工智能的一个分支,已在众多领域得到成功运用,把它用在教学领域还是一个新的尝试．结合多元智能理论,以学生的多元智能光谱为处理对象,建立了一个BP神经网络,利用LM算法和MATLAB语言,通过网络训练,得到了一种学生素质的处理方法．     

Neural decoding based on probabilistic neural network

Brain-machine interface (BMI) has been developed due to its possibility to cure severe body paralysis. This technology has been used to realize the direct control of prosthetic devices,such as robot arms,computer cursors,and paralyzed muscles. A variety of neural decoding algorithms have been designed to explore relationships between neural activities and movements of the limbs. In this paper,two novel neural decoding methods based on probabilistic neural network (PNN) in rats were introduced,the PNN decode...     

概率神经网络在汽轮发电机组故障诊断中的应用

汽轮发电机组逐渐智能化,功能不断增强,但不确定性因素和不确定性信息仍然大量存在。针对该问题,用概率神经网络（PNN）诊断汽轮发电机组故障。PNN优点较多,机器学习算法简易、方便训练,相比于传统样本处理方法,PNN可训练样本并引入训练网络,更好地确保诊断结果正确率与可信度。MATLAB仿真结果表明,PNN在保证诊断结果准确的基础上,速度更快、分类性能大幅提高,诊断效率也提高至98%。     

Explaining student success in engineering education at Delft University of Technology: a literature synthesis

Student success is among the most widely researched areas in tertiary education. Generalisability of research in this field is problematic due to cultural and structural differences between countries, institutions and programmes where the research is done. Engineering education in the Netherlands has not been studied in depth. In this paper, outcomes of studies done outside and inside engineering and outside and inside the Netherlands are discussed to help understand the complexity of student retention issues. Although generalisation is an issue, there are a number of concepts and variables that surface in many of these studies, including students’ background and disposition variables, education attributes, variables concerning educational climate and student behaviour. How these variables are related and how a university can apply the outcomes of research in this field of study are discussed in this paper.     

基于BP神经网络的公共建筑用电能耗预测研究

数据挖掘技术与建筑工程之间的知识跨度较大,将两者完美融合起来存在一定难度,实际工程中也缺乏相关案例,因此在建筑工程中应用数据挖掘技术挖掘相关信息,为大型公共建筑用电能耗预测提供参考依据,是建筑领域一种新的发展趋势。基于对公共建筑用电能耗特性的分析,可利用Python构建BP神经网络建筑能耗预测模型,再将某公共建筑作为研究对象,确定影响建筑用电能耗的关键因素,并将其作为网络的输入参数进行学习训练得出预测值。研究结果表明：预测模型在不同参数下,当隐含层个数为8时,误差平方和最小,为0.000 139 6,此时BP神经网络能够较精确地预测公共建筑用电能耗值。     

基于Elman神经网络的商品销售额建模及预测

为提高预测精度,采用Elman神经网络对商品销售额进行预测.在分析样本数据特性基础上,采用Elman神经网络对商品销售额的预测进行建模,通过实例验证模型的有效性并对比分析BP神经网络的预测效果.研究结果表明,基于Elman神经网络的商品销售额预测能够较好地反映商品销售额的变化趋势且提高了预测精度,验证了该方法在商品销售额预测中的可行性.     

基于粗糙集-LVQ神经网络的入侵检测模型研究

提出了一种基于粗糙集-LVQ神经网络的入侵检测模型。首先将粗糙集作为前置处理系统对属性进行约简,再根据粗糙集处理后的学习样本集,构建LVQ神经网络结构,实现对入侵进行检测。实验结果表明该方法优于其他同类方法。     

Retention in a Bachelor of Education (Early childhood studies) course: students say why they stay and others leave

The literature suggests that student attrition at the university level has been of growing concern in many countries. Student attrition has a number of implications for universities, chief amongst them are losses to revenue and investment in higher education. While many studies have examined causes for attrition from an institutional perspective, this study examines how the Bachelor of Education (Early childhood studies) that sits within the School of Education can support the retention of students from the students’ perspectives. Using a qualitative methodology that recorded up to 40 hours of interviews with 20 students provided insights into why they stay at university and what aspects of the Early childhood studies degree support them in staying. Conversely, the students also comment on those aspects that are not supportive and could impact on retention. The data revealed that a sense of belonging was very important and that particular course-specific activities at certain times supported developing this belonging. The students identified the importance of a mentorship program and how this program needs to be made available from the commencement of their course. Other factors contributing to retention included well-placed practicums, career choice, designated study times, as well as study groups and supportive and responsive lecturers.     

基于粗糙集与RBF神经网络的农业总产值预测方法

提出了把粗糙集和R BF神经网络相结合应用于农业总产值预测的方法。首先用粗糙集对影响农业总产值的多个因素进行属性约简,选择主要影响因素,去除冗余信息;然后利用RBF神经网络建立预测模型。最后对该模型的预测结果与因子分析神经网络模型的预测结果进行了比较,表明了该模型的有效性和优越性。     

融合Elman神经网络与EMD的交通拥堵指数预测

为了提高传统方法生成交通拥堵指数（TPI）的准确率,引入一种基于经验模态分解（EMD）与Elman神经网络的组合模型实现交通拥堵指数预测。首先,利用EMD将TPI序列分解为不同时间尺度下的IMF分量和剩余分量;然后,通过偏自相关函数（PACF）计算各分量的滞后期数,以此确定各分量在Elman神经网络中的输入和输出变量;之后,通过上述方法计算出各分量预测值并相加;最后,计算出总预测结果。通过计算结果可知,EMD-PACF-Elman预测方法3个评价指标（平均绝对误差、均方误差、平均绝对百分误差）的计算结果与单一Elman神经网络模型、EMD-Elman神经网络模型、单一BP神经网络模型、EMD-BP神经网络模型相比都为最低,分别为0.562 4、0.598 9、0.110 7。因此, EMD-PACF-Elman预测方法可以有效地预测TPI,同时也为进一步预测交通拥堵趋势提供了依据。     

基于BP神经网络的认知诊断评估分类准确率影响因素分析

BP神经网络是目前应用最广泛的人工神经网络模型之一,在分类和识别上表现出良好的特性,因此被研究者用于认知诊断评估以对被试进行诊断分类。通过模拟研究,考查属性个数、属性层级关系、测验长度、题目质量、测试样本量5个因素对BP神经网络在认知诊断中分类准确性的影响。结果表明:1)基于BP神经网络的认知诊断分类准确率不依赖于测试样本量;2)题目质量和测验长度对BP神经网络的诊断准确率有显著的积极影响;3)属性个数对BP神经网络的分类准确率有消极影响;4)题目质量一定程度上会影响BP诊断方法在不同属性层级结构上的分类准确率。     

基于组合神经网络模型的新冠疫情传播预测分析

在全球抗击新型冠状病毒肺炎（COVID-19）疫情的过程中,合理的疫情传播预测对于疫情防控有重要参考意义。为了对病毒传播进行合理预测,针对传统疫情传播预测模型存在的不足,提出一种组合式神经网络的疫情传播预测模型,并将其应用于湖北省1月29日-3月15日每日新增确诊人数预测及湖北省每日累计确诊人数预测。预测结果分析显示,该神经网络预测模型预测结果可靠有效。模型性能分析结果表明,组合式神经网络预测模型平均相对误差（MRE）不超过0.16,均方误差（MSE）不超过0.1,均方根误差（RMSE）为0.262 9,性能明显优于其它几种神经网络预测模型。基于武汉市与广东省疫情传播预测的实证结果显示模型具有较好的适用性及准确性。     

Creating a connection: tackling student attrition through curriculum development

The issue of student attrition (where students leave a programme of study for any reason) is maintaining a high profile across the higher education sector and is a key concern for those delivering nurse education. Many authors have pointed out the cause for concern that student nurse attrition raises and its impact on the higher education institution (HEI), on the student body generally, on partners in practice, and on the student who leaves. The reasons that student nurses leave their programmes of study are similar to the reasons that other students leave (financial, personal, academic, lack of support, etc). For student nurses there are also issues with their experiences in practice. It is clear that with all student attrition, the reasons for leaving are complex in nature. It must be accepted that there are some issues that cannot be addressed by HEIs, but that ways to tackle attrition and to enhance student retention must be high on the agenda of any education provider. Individual programmes of study can utilise approaches to curriculum development that aim to enhance the student experience. The introduction of an enquiry‐based learning strand to the curriculum of one pre‐registration nursing programme aims to address some of the reasons for student attrition, namely: lack of support, poor academic performance, lack of study skills, disillusionment and low motivation, lack of confidence, and the theory‐practice gap resulting in stressful practice experiences. Enquiry‐based learning (EBL) is not seen as a panacea for student attrition, but it is seen as a positive way in which to enhance the student experience.     

College Counseling and Student Retention:

Although it is commonly believed that college counseling positively affects student retention and further, that there are convincing data and evidence to support such a belief, an extensive literature search found no objective evidence at all on the relationship of a counseling center's conventional one-to-one counseling and therapy activities with short-term student retention. Two program evaluations which attempt to relate one-to-one individual counseling and psychotherapy to student retention rates at an urban regional state university are discussed. Changes in attendance rates, such as dropping out or retuning between quarters are the criteria used to evaluate program effects for both studies. The first study searched for discernible statistically significant effects. Three-quarters of comparisons of upperclassmen found conceivable retention advantages of students receiving counseling center services over a base-rate group. Yet the freshmen results suggest that counseling may be associates with student attrition, not retention. The second study investigated the relationship between retention and possibly relevant presenting problems and demographic variables of students undergoing individual counseling. No attendance differences among counseled students were found due to age, gender, or presenting problem. Following a discussion of the results, suggentions for future research are presented. A longitudinal, quasi-experimental controlled study of the impact counseling on student retention is clearly needed.     

Considering student retention as a complex system: a possible way forward for enhancing student retention

This study uses multilayer minimum spanning tree analysis to develop a model for student retention from a complex system perspective, using data obtained from first-year engineering students at a large well-regarded institution in the European Union. The results show that the elements of the system of student retention are related to one another through a network of links and that some of these links were found to be strongly persistent across different scales (group sizes). The links were also seen to group together in different clusters of strongly related elements. Links between elements across a wide range of these clusters would have system-wide influence. It was found that there were no elements that are both persistent and have system-wide effects. This complex system view of student retention explains why actions to enhance student retention aimed at single elements in the system have had such limited impact. This study therefore points to the need for a more system-wide approach to enhancing student retention.     

基于 WOA-GRU 的销售预测研究

防晒用品属于快速消费品,其销售情况受季节变动影响,具有非线性特征。人工神经网络强大的学习能力使其在销售量预测领域显示出非线性预测的应用优势。作为新兴智能算法,鲸鱼群算法具有快速收敛、全局优化等特点。为精准预测上海某化妆品公司防晒产品销量,基于鲸鱼算法,构建优化网络权值的门控循环神经网络模型。实验结果表明,经鲸鱼算法优化后的门控循环神经网络训练相对误差比优化前降低 0.6 个百分点,测试相对误差比优化前降低约 2 个百分点。     

提高多层前向神经网络泛化能力的讨论

泛化能力是多层前向神经网络最重要的性能．泛化问题已成为目前神经网络领域的研究热点。本文从泛化理论现有提高神经网络泛化能力的方法等几个方面总结了当前神经网络结构优化与泛化能力研究的现状。神经网络泛化能力的提高可通过神经网络结构的优化和正则化等方法加以实现，并对提高网络泛化能力问题进行讨论。     

一种神经网络模型在混凝土配比设计中的应用

本文描述了人工神经网络的模型和算法，给出了BP神经网络模型和算法的相关变量的处理方法。用这种方法，对混凝土配合比试验数据进行分析预测，效果良好。说明BP神经网络模型可用于混凝土混合比强度预测中。     

Peer Coaching of At-Risk Students in a Remedial Mathematics Course

Remedial mathematics at the college level has evidenced strikingly low retention and passing rates. It has been shown that at least 25% of the variation in student performance is explained by students’ affective variables such as attitudes, study skills, and mathematics anxiety. This study endeavored to address affective variables by: (a) incorporating the teaching of study skills into an elementary algebra course; and (b) identifying at-risk students and assigning these students “coaches,” who functioned both as tutors and counselors, providing regular personalized assistance. The results showed that the attrition rate was significantly lower in the treatment groups, and the overall passing rate for coached students was significantly higher.