lower accuracy
Recently Published Documents





2022 ◽  
Melek Tassoker ◽  
Muhammet Usame Ozic ◽  
Fatma Yuce

Abstract Objective: The aim of the present study was to predict osteoporosis on panoramic radiographs of women over 50 years of age through deep learning algorithms.Method: Panoramic radiographs of 744 female patients over 50 years of age were labeled as C1, C2, and C3 depending on mandibular cortical index (MCI). According to this index; C1: presence of a smooth and sharp mandibular cortex (normal); C2: resorption cavities at endosteal margin and 1 to 3-layer stratification (osteopenia); C3: completely porotic cortex (osteoporosis). The data of the present study were reviewed in different categories including C1-C2-C3, C1-C2, C1-C3 and C1-(C2+C3) as two-class and three-class prediction. The data were separated as 20% random test data; and the remaining data were used for training and validation with 5-fold cross-validation. AlexNET, GoogleNET, ResNET-50, SqueezeNET, and ShuffleNET deep learning models are trained through the transfer learning method. The results were evaluated by performance criteria including accuracy, sensitivity, specificity, F1-score, AUC and training duration. Findings: The dataset C1-C2-C3 has an accuracy rate of 81.14% with AlexNET; the dataset C1-C2 has an accuracy rate of 88.94% with GoogleNET; the dataset C1-C3 has an accuracy rate of 98.56% with AlexNET; and the dataset C1-(C2+C3) has an accuracy rate of 92.79% with GoogleNET. Conclusion: The highest accuracy was obtained in differentiation of C3 and C1 where osseous structure characteristics change significantly. Since the C2 score represent the intermediate stage (osteopenia), structural characteristics of the bone present behaviors closer to C1 and C3 scores. Therefore, the data set including the C2 score provided relatively lower accuracy results.

Houjie Li ◽  
Min Yang ◽  
Yu Zhou ◽  
Ruirui Zheng ◽  
Wenpeng Liu ◽  

Partial label learning is a new weak- ly supervised learning framework. In this frame- work, the real category label of a training sample is usually concealed in a set of candidate labels, which will lead to lower accuracy of learning al- gorithms compared with traditional strong super- vised cases. Recently, it has been found that met- ric learning technology can be used to improve the accuracy of partial label learning algorithm- s. However, because it is difficult to ascertain similar pairs from training samples, at present there are few metric learning algorithms for par- tial label learning framework. In view of this, this paper proposes a similar pair-free partial la- bel metric learning algorithm. The main idea of the algorithm is to define two probability distri- butions on the training samples, i.e., the proba- bility distribution determined by the distance of sample pairs and the probability distribution de- termined by the similarity of candidate label set of sample pairs, and then the metric matrix is ob- tained via minimizing the KL divergence of the two probability distributions. The experimental results on several real-world partial label dataset- s show that the proposed algorithm can improve the accuracy of k-nearest neighbor partial label learning algorithm (PL-KNN) better than the ex- isting partial label metric learning algorithms, up to 8 percentage points.

2022 ◽  
Vol 12 (2) ◽  
pp. 584
Sherif M. Hanafy

Near-surface high-resolution seismic mapping is very important in many applications such as engineering and environmental. However, the conventional setup of the seismic technique requires planting geophones, connecting cables, and then collecting all equipment after completing the survey, which is time-consuming. In this study, we suggest using a land-streamer setup rather than the conventional setup for fast, accurate, and high-resolution near-surface seismic surveys. Only one field data set is recorded using both the conventional and the land-streamer setups. The recorded data is then compared in terms of time, frequency, wavenumber domains, and acquisition time. Following this, we compared the accuracy of the subsurface mapping of both setups using a synthetic example. The results show that the conventional setup can reach deeper depths but with lower accuracy, where the errors in imaging the local anomalies’ widths and thicknesses are 77% to 145% and 35% to 50%, respectively. The land-streamer setup provides accurate near-surface results but shallower penetration depth, here the errors in the anomalies’ widths and thicknesses are 5% to 12% and 10% to 20%, respectively.

2022 ◽  
Vol 16 (1) ◽  
pp. 0-0

Recent advances in machine learning have shown promising results for detecting network intrusion through supervised machine learning. However, such techniques are ineffective for new types of attacks. In the preferred unsupervised and semi-supervised cases, these newer techniques suffer from lower accuracy and higher rates of false alarms. This work proposes a machine learning model that combines auto-encoder with one-class support vectors machine. In this model, the auto-encoders learn the representation of the input data in a latent space and reduces the dimensionality of the input data. The dimensionality-reduced input is then extracted from the auto-encoder and passed to a one-class support vectors machine to classify the network event as an attack or a normal event. The model is trained on normal network events only. The proposed model is then evaluated and compared with several existing models. It achieves high accuracy when tested on the NSL-KDD and KDD99 datasets, with total accuracies of 96.24% and 99.45%, respectively.

2021 ◽  
Vol 26 (4) ◽  
pp. 847-862
Suji Kim ◽  
Jee Eun Sung

Objectives: The purpose of this study is to investigate how aging influences sentence processing when noun-phrases are presented differently.Methods: A total of 40 participants participated in the study ranging in age from 19 to 71. All were presented with sentences and pictures under either dative or accusative conditions. After that, they were asked to judge if the sentences were correct or incorrect.Results: First, there were significant differences between the older adults and younger adults in accuracy. The older group showed lower accuracy in the sentence judgment task. Second, there were significant differences between the older adults and younger adults in response time. The older group needed more time due to their lower cognitive resources. They made more errors when accusative noun phrases were provided. Third, the fixation proportion of the target stimulus between regions were significant in both types of dative and accusative noun phrase presentation. The older group showed lower proportions in the last region of the sentence.Conclusion: These results shows that both the elderly and the young gradually deal with the meaning of words through the noun phrase information. However, the elderly showed difficulty in assigning the correct thematic roles by using case-markers, given the lower proportion of fixation in the region where the target stimuli are presented. It is expected that difficulties in the communication process of the elderly will be better understood through this study.

2021 ◽  
Jihong Zhang ◽  
Terry Ackerman ◽  
Yurou Wang

Fitting item response theory (IRT) models using the generalized mixed logistic regression model (GLMM) has become more popular in large-scale assessment because GLMM allows combining complicated multilevel structures (i.e., students are nested in classrooms which are nested in schools) with IRT measurement models. However, the estimation accuracy of item parameters between these two models is not well examined. This study aimed to compare the estimation results of the GLMM based 2PL model (using the PLmixed R package) with the traditional IRT model (using flexMIRT software) under different sample sizes (N= 500, 1000, 5000) and test length (J = 15, 21) conditions. The simulation results showed that for both the GLMM-based method and the traditional method, item threshold estimates had lower bias than item discrimination parameters. We also found that according to the simulation study, GLMM estimates via PLmixed had lower accuracy than traditional IRT modeling via flexMIRT for items with high discrimination.

Complexity ◽  
2021 ◽  
Vol 2021 ◽  
pp. 1-11
Jing Liu ◽  
Junfang Guo ◽  
Qi Li

Community structure is one of the most important characteristics of complex networks, which has important applications in sociology, biology, and computer science. The community detection method based on local expansion is one of the most adaptable overlapping community detection algorithms. However, due to the lack of effective seed selection and community optimization methods, the algorithm often gets community results with lower accuracy. In order to solve these problems, we propose a seed selection algorithm of fusion degree and clustering coefficient. The method calculates the weight value corresponding to degree and clustering coefficient by entropy weight method and then calculates the weight factor of nodes as the seed node selection order. Based on the seed selection algorithm, we design a local expansion strategy, which uses the strategy of optimizing adaptive function to expand the community. Finally, community merging and isolated node adjustment strategies are adopted to obtain the final community. Experimental results show that the proposed algorithm can achieve better community partitioning results than other state-of-the-art algorithms.

Methane ◽  
2021 ◽  
Vol 1 (1) ◽  
pp. 38-57
Diana Sorg

The handheld, portable laser methane detector (LMD) was developed to detect gas leaks in industry from a safe distance. Since 2009, it has also been used to measure the methane (CH4) concentration in the breath of cattle, sheep, and goats to quantify their CH4 emissions. As there is no consensus on a uniform measurement and data-analysis protocol with the LMD, this article discusses important aspects of the measurement, the data analysis, and the applications of the LMD based on the literature. These aspects, such as the distance to the animal or the activity of the animals, should be fixed for all measurements of an experiment, and if this is not possible, they should at least be documented and considered as fixed effects in the statistical analysis. Important steps in data processing are thorough quality control and reduction in records to a single point measurement or “phenotype” for later analysis. The LMD can be used to rank animals according to their CH4 breath concentration and to compare average CH4 production at the group level. This makes it suitable for genetic and nutritional studies and for characterising different breeds and husbandry systems. The limitations are the lower accuracy compared to other methods, as only CH4 concentration and not flux can be measured, and the high amount of work required for the measurement. However, due to its flexibility and non-invasiveness, the LMD can be an alternative in environments where other methods are not suitable or a complement to other methods. It would improve the applicability of the LMD method if there were a common protocol for measurement and data analysis developed jointly by a group of researchers.

2021 ◽  
Vol 13 (3) ◽  
Donald E Brannen ◽  
Melissa Howell ◽  
Ashley Steveley ◽  
Jeff Webb ◽  
Deidre Owsley

Background:Fall injuries (FI) are a priority for public health planning. Syndromic surveillance (SS) is used to detect outbreaks, environmental exposures, and bioterrorism in real time. Since information is gathered on patients, the utility of using this system for FI should be evaluated. Methods:Strategies to integrate FI medical and SS data were compared using a cohort versus case control (CC) study design. Results:The CC study was accurate 77.7% (57.7-91.3) of the time versus 100% for a cohort design. The CC study design found FI increased for older age groups, female gender, November, and December months. Dates with any freezing temperature had a higher case fatality rate. Repeat acute care visits increased the risk of FI diagnosis by over 6% and trended upward with each visit (R=.333, p<.001). Conclusions:The CC diagnostic quality of FI were better for age and gender than for area. The CC study found the indicators of increased risk of FI including: Freezing temperature, repeat acute care visits, older age groups, female gender, November, and December months. A gradient of increasing odds of FI with the number of acute care visits provides proof that community fall prevention programs should focus on those most likely to fall. A CC design of SS data can quickly identify indicators of FI with a lower accuracy but with less cost than a full cohort study, thus providing a method to focus local public health interventions.

Sign in / Sign up

Export Citation Format

Share Document