Hierarchical Discriminant Analysis Research Articles

The analysis of intelligent real-time image recognition technology based on mobile edge computing and deep learning

This article aims to improve the accuracy of real-time image recognition in the context of the Internet of Things (IoT), reduce the core network pressure of the IoT and the proportion of IoT broadband, and meet people’s demand for internet image transmission. An intelligent image fusion system based on mobile edge computing (MEC) and deep learning is proposed, which can extract the features of images and optimize the sum of intra-class distance and inter-class distance relying on the hierarchical mode of deep learning, and realize distributed computing with the edge server and base station. Through comparison with other algorithms and strategies on the text and character data sets, the effectiveness of the constructed system is verified in the performance of the algorithm and the IoT. The results reveal that the application of the unsupervised learning hierarchical discriminant analysis (HDA) has better accuracy and recall in various databases compared with conventional image recognition algorithms. When the sum intra-class and inter-class distance K is 2, the accuracy of the algorithm can be as high as 98%. The combination of MEC and layered algorithms effectively reduces the pressures of core network and shortens the response time, greatly reduces the broadband occupancy ratio. The performance of IoT is increased by 37.03% compared with the general extraction and common cloud computing. Image recognition based on the MEC architecture can reduce the amount of network transmission and reduce the response time under the premise of ensuring the recognition rate, which can provide a theoretical basis for the research and application of user image recognition under the IoT.

Adoption of image surface parameters under moving edge computing in the construction of mountain fire warning method.

In order to cope with the problems of high frequency and multiple causes of mountain fires, it is very important to adopt appropriate technologies to monitor and warn mountain fires through a few surface parameters. At the same time, the existing mobile terminal equipment is insufficient in image processing and storage capacity, and the energy consumption is high in the data transmission process, which requires calculation unloading. For this circumstance, first, a hierarchical discriminant analysis algorithm based on image feature extraction is introduced, and the image acquisition software in the mobile edge computing environment in the android system is designed and installed. Based on the remote sensing data, the land surface parameters of mountain fire are obtained, and the application of image recognition optimization algorithm in the mobile edge computing (MEC) environment is realized to solve the problem of transmission delay caused by traditional mobile cloud computing (MCC). Then, according to the forest fire sensitivity index, a forest fire early warning model based on MEC is designed. Finally, the image recognition response time and bandwidth consumption of the algorithm are studied, and the occurrence probability of mountain fire in Muli county, Liangshan prefecture, Sichuan is predicted. The results show that, compared with the MCC architecture, the algorithm presented in this study has shorter recognition and response time to different images in WiFi network environment; compared with MCC, MEC architecture can identify close users and transmit less data, which can effectively reduce the bandwidth pressure of the network. In most areas of Muli county, Liangshan prefecture, the probability of mountain fire is relatively low, the probability of mountain fire caused by non-surface environment is about 8 times that of the surface environment, and the influence of non-surface environment in the period of high incidence of mountain fire is lower than that in the period of low incidence. In conclusion, the surface parameters of MEC can be used to effectively predict the mountain fire and provide preventive measures in time.

Hierarchical Discriminant Analysis.

The Internet of Things (IoT) generates lots of high-dimensional sensor intelligent data. The processing of high-dimensional data (e.g., data visualization and data classification) is very difficult, so it requires excellent subspace learning algorithms to learn a latent subspace to preserve the intrinsic structure of the high-dimensional data, and abandon the least useful information in the subsequent processing. In this context, many subspace learning algorithms have been presented. However, in the process of transforming the high-dimensional data into the low-dimensional space, the huge difference between the sum of inter-class distance and the sum of intra-class distance for distinct data may cause a bias problem. That means that the impact of intra-class distance is overwhelmed. To address this problem, we propose a novel algorithm called Hierarchical Discriminant Analysis (HDA). It minimizes the sum of intra-class distance first, and then maximizes the sum of inter-class distance. This proposed method balances the bias from the inter-class and that from the intra-class to achieve better performance. Extensive experiments are conducted on several benchmark face datasets. The results reveal that HDA obtains better performance than other dimensionality reduction algorithms.

Hierarchical Discriminant Analysis and Its Application

This article presents a new discriminant analysis method and tries to improve the accuracy of discriminant problems without reducing variables. This method establishes two layers of discriminant functions. For the first layer, a linear discriminant function is constructed to determine the sample type. Those samples’ type are not determined at the first layer, they will be distinguished as sub-samples, then a second layer of a nonlinear discriminant function will be established to determine the sub-sample type. The examples show that this method can handle discriminant problems of complex data, and results of this method are better than some common methods.

A Comparison of Doubly Hierarchical Discriminant Analyses for Multiple Class Longitudinal Data from EEG Experiments

This paper proposes a general and simple procedure that can be applied to establish classification rules for application to multiple-class longitudinal data. The procedure is applied to preclinical pharmaco-electroencephalogram (EEG) studies aiming at characterizing psychotropic drug effects on the basis of spectral EEG analysis. It is a flexible hierarchical supervised learning tool that takes into account the specific nature of the multiple drug classes, as well as the longitudinal aspect of the data. Several variations of this procedure are applied to the EEG data, generally producing comparable results, in particular similar association between the sleeping stages and the psychotropic drug classes.

Psychotropic Drug Classification Based on Sleep–Wake Behaviour of Rats

Summary The aim of the paper is to present methodology for the classification of potential psychotropic drugs on the basis of their activity. We first sketch the background of this class of drugs and then zoom in on so-called pharmacoelectroencephalogram studies. These data pose some statistical challenges. For classification purposes, we propose a flexible hierarchical discriminant analysis tool, allowing us to take the specific nature of the drug class into account, as well as the features of the mixed models, in combination with fractional polynomials, fitted to the electroencephalogram data. The method is evaluated against the background of existing methods. The method's performance is studied by using a comprehensive analysis of a large electroencephalogram data set.

Texture Classification Using Hierarchical Linear Discriminant Space

As a representative of the linear discriminant analysis, the Fisher method is most widely used in practice and it is very effective in two-class classification. However, when it is expanded to a multi-class classification problem, the precision of its discrimination may become worse. A main reason is an occurrence of overlapped distributions on the discriminant space built by Fisher criterion. In order to take such overlaps among classes into consideration, our approach builds a new discriminant space by hierarchically classifying the overlapped classes. In this paper, we propose a new hierarchical discriminant analysis for texture classification. We divide the discriminant space into subspaces by recursively grouping the overlapped classes. In the experiment, texture images from many classes are classified based on the proposed method. We show the outstanding result compared with the conventional Fisher method.

A Hierarchical Discriminant Analysis for Species Identification in Raw Meat by Visible and near Infrared Spectroscopy

The problem tackled is that of discriminating between chicken, turkey, pork, beef and lamb using visible and near infrared spectra collected from homogenised meat samples. This five-group classification task was treated as a hierarchical sequence of binary splits: white versus red meat, then either poultry versus pork or beef versus lamb, and finally, in the case of poultry, chicken versus turkey. Most of the splits were achieved by linear discriminant analysis applied to principal component scores. The chicken versus turkey split worked better with soft independent modelling of class analogy (SIMCA). One of the attractions of the hierarchical approach is that different methods can be used for different splits. The results, with only two classification errors on 115 validation samples, were better than those achieved in previous analyses of the same data.

Does choosing a treatment depend on making a diagnosis? US and French physicians' decision making about acute otitis media.

The classic sequential processing model of clinical decision making-in which the treatment choice follows and depends on the diagnostic judgment-may in some cases be replaced by a processing model in which the treatment choice depends on an independent assessment of the diagnostic and other cues. The aim of this study was to determine which processing model would better describe physicians' treatment choices in a simulated clinical task. Seventy-five US and French primary care physicians were presented twice, in a different order, with the same set of 46 scenarios of 15-month-old children suspected of having acute otitis media (AOM). They rated in one set the probability of AOM and in the other set whether they would treat the child with antibiotics (and how confident they felt in their decision). Linear regression analyses revealed the individuals' 2 judgment policies. Hierarchical discriminant analysis was used to analyze the variance explained in the treatment choice by, 1st, the diagnostic judgment, 2nd, the cues specific to treatment, and 3rd, the cues specific to diagnosis. Even when choosing treatment, the participants placed greatest weight on diagnostic cues, especially the ear findings. Only 28% used the cues that reflected parental issues. For 36%, the diagnostic cues had an effect on the treatment choice independent of the effect (if any) of the diagnostic judgment. In deciding how to treat AOM, the majority of the participating US and French primary care physicians followed the classic sequential processing model, but a substantial minority used instead an independent processing model.

Assessing adaptive and maladaptive coping in the early phase of acute myocardial infarction

We assessed retrospectively symptom management behaviors among 45 patients with acute myocardial infarction (AMI). Cluster analysis was conducted on data from the Structured Interview (SI), in order to group patients according to the global TABP, Anger-In, Potential for Hostility, and a newly derived measure of Behavioral Compensation for Stress (BEH-COMP). Two groups were identified. MALADAPT SI-copers (N = 24) had higher scores on Anger-In and Hostility and lower scores on BEH-COMP in comparison to the ADAPT SI-copers (N = 21), while no difference was observed on TABP. The SI groups were next compared on AMI symptom management behaviors. A hierarchical discriminant analysis found that the MALADAPT group reported greater distraction from AMI symptoms, were relief-seeking behavior, and greater perceived vulnerability to reinfarction. AMI coping behaviors correlated meaningfully with delay in seeking medical assistance. Further research is warranted, given the potential for using the ADAPT/MALADAPT SI-profiles to predict adjustment to AMI.

Use of mental health services among victims of crime: Frequency, correlates, and subsequent recovery.

The use of mental health services was examined within a sample of 392 victims of crime who were generally representative of that population in Kentucky. Respondents were interviewed twice at a 6-month interval. Of these victims, 12% had contact with mental health professionals within the first few months postcrime. Hierarchical discriminant analyses revealed that use of mental health services was most prevalent when depressive symptoms were present and the crime involved violence. Among victims of violence, urban residence, high social support, internal locus of control, and prior crime experience were also associated with use. The victims who recovered from the event most rapidly were those who reported receiving mental health services at both Waves 1 and 2. Implications of the study for policies and programs are discussed.

Correlates of Pregnancy and Sexual Experience in Australian Adolescents

In 1983 a longitudinal study was commenced in the Hunter Region of New South Wales to investigate the development of adolescent psychosocial morbidity. This paper reports analyses based on data obtained during the screening phase of that project, in which 2157 adolescents completed a Social and Emotional Development questionnaire. Those surveyed comprised approximately 25% of the 14-16 year olds in the region.Adolescents were allocated to 1 of 3 groups on the basis of their sexual experience: no sexual experience (26%), minor sexual experience (such as petting, 47%), and major sexual experience (sexual intercourse on at least one occasion, 27%). Overall, 31% of the males and 24% of the females were in the major sexual experience group; 1 in 15 of the females in this group (n = 17) reported that she had been pregnant. Just over half (52%) of the girls in the major sexual experience group had used contraceptives, compared with 42% of the boys.A series of hierarchical discriminant analyses were undertaken to...