Statistical Method and Deep Learning Model for Sea Surface Temperature Prediction (수온 데이터 예측 연구를 위한 통계적 방법과 딥러닝 모델 적용 연구)
-
- Journal of the Korean Society of Marine Environment & Safety
- /
- v.29 no.6
- /
- pp.543-551
- /
- 2023
As climate change continues to prompt an increasing demand for advancements in disaster and safety management technologies to address abnormal high water temperatures, typhoons, floods, and droughts, sea surface temperature has emerged as a pivotal factor for swiftly assessing the impacts of summer harmful algal blooms in the seas surrounding Korean Peninsula and the formation and dissipation of cold water along the East Coast of Korea. Therefore, this study sought to gauge predictive performance by leveraging statistical methods and deep learning algorithms to harness sea surface temperature data effectively for marine anomaly research. The sea surface temperature data employed in the predictions spans from 2018 to 2022 and originates from the Heuksando Tidal Observatory. Both traditional statistical ARIMA methods and advanced deep learning models, including long short-term memory (LSTM) and gated recurrent unit (GRU), were employed. Furthermore, prediction performance was evaluated using the attention LSTM technique. The technique integrated an attention mechanism into the sequence-to-sequence (s2s), further augmenting the performance of LSTM. The results showed that the attention LSTM model outperformed the other models, signifying its superior predictive performance. Additionally, fine-tuning hyperparameters can improve sea surface temperature performance.
Recently, stratigraphic reservoirs are getting more attention than structural reservoirs which have mostly developed. However, recognizing stratigraphic thin gas reservoirs in a stacked section is usually difficult because of tuning effects. Moreover, if the reflections from the brine-saturated region of a thin layer have the same polarity with those from the gas-saturated region, we could not easily identify the gas reservoir with conventional data processing technique. In this study, we introduced a way to delineate the gas-saturated region in a thin layer reservoir using a spectral decomposition method. First of all, amplitude spectrum with the variation of the frequency and the incident angle was investigated for the medium which represents property of Class 3, Class 1 or Class 4 AVO response. The results show that the maximum difference in the amplitude spectra between brine and gas-saturated thin layers occurs around the peak frequency independent of the incident angle and the type of AVO responses. In addition, the amplitude spectra of the gas-saturated zone are greater than those of brine-saturated one in Class 3 and Class 4 at the peak frequency while those of phenomenon occur oppositely in Class 1. Based on the results, we applied spectral decomposition method to the stacked section in order to distinguish the gas-saturated zone from the brine-saturated zone in a thin layer reservoir. To verify our new method, we constructed a thin-layer velocity model which contains both gas and brine-saturated zones which have the same reflection polarities. As a result, in the spectral decomposed sections near the peak frequency obtained by Wigner-Ville Distribution (WVD), we could identify the difference between reflections from gas- and brinesaturated region in the thin layer reservoir, which was hardly distinguishable in the stacked section.
Rainfall-runoff prediction studies using deep learning while considering catchment attributes have been gaining attention. In this study, we selected two models: the Transformer model, which is suitable for large-scale data training through the self-attention mechanism, and the LSTM-based multi-state-vector sequence-to-sequence (LSTM-MSV-S2S) model with an encoder-decoder structure. These models were constructed to incorporate catchment attributes and predict the inflow of 10 multi-purpose dam watersheds in South Korea. The experimental design consisted of three training methods: Single-basin Training (ST), Pretraining (PT), and Pretraining-Finetuning (PT-FT). The input data for the models included 10 selected watershed attributes along with meteorological data. The inflow prediction performance was compared based on the training methods. The results showed that the Transformer model outperformed the LSTM-MSV-S2S model when using the PT and PT-FT methods, with the PT-FT method yielding the highest performance. The LSTM-MSV-S2S model showed better performance than the Transformer when using the ST method; however, it showed lower performance when using the PT and PT-FT methods. Additionally, the embedding layer activation vectors and raw catchment attributes were used to cluster watersheds and analyze whether the models learned the similarities between them. The Transformer model demonstrated improved performance among watersheds with similar activation vectors, proving that utilizing information from other pre-trained watersheds enhances the prediction performance. This study compared the suitable models and training methods for each multi-purpose dam and highlighted the necessity of constructing deep learning models using PT and PT-FT methods for domestic watersheds. Furthermore, the results confirmed that the Transformer model outperforms the LSTM-MSV-S2S model when applying PT and PT-FT methods.
Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.
Recently, as deep learning has attracted attention, the use of deep learning is being considered as a method for solving problems in various fields. In particular, deep learning is known to have excellent performance when applied to applying unstructured data such as text, sound and images, and many studies have proven its effectiveness. Owing to the remarkable development of text and image deep learning technology, interests in image captioning technology and its application is rapidly increasing. Image captioning is a technique that automatically generates relevant captions for a given image by handling both image comprehension and text generation simultaneously. In spite of the high entry barrier of image captioning that analysts should be able to process both image and text data, image captioning has established itself as one of the key fields in the A.I. research owing to its various applicability. In addition, many researches have been conducted to improve the performance of image captioning in various aspects. Recent researches attempt to create advanced captions that can not only describe an image accurately, but also convey the information contained in the image more sophisticatedly. Despite many recent efforts to improve the performance of image captioning, it is difficult to find any researches to interpret images from the perspective of domain experts in each field not from the perspective of the general public. Even for the same image, the part of interests may differ according to the professional field of the person who has encountered the image. Moreover, the way of interpreting and expressing the image also differs according to the level of expertise. The public tends to recognize the image from a holistic and general perspective, that is, from the perspective of identifying the image's constituent objects and their relationships. On the contrary, the domain experts tend to recognize the image by focusing on some specific elements necessary to interpret the given image based on their expertise. It implies that meaningful parts of an image are mutually different depending on viewers' perspective even for the same image. So, image captioning needs to implement this phenomenon. Therefore, in this study, we propose a method to generate captions specialized in each domain for the image by utilizing the expertise of experts in the corresponding domain. Specifically, after performing pre-training on a large amount of general data, the expertise in the field is transplanted through transfer-learning with a small amount of expertise data. However, simple adaption of transfer learning using expertise data may invoke another type of problems. Simultaneous learning with captions of various characteristics may invoke so-called 'inter-observation interference' problem, which make it difficult to perform pure learning of each characteristic point of view. For learning with vast amount of data, most of this interference is self-purified and has little impact on learning results. On the contrary, in the case of fine-tuning where learning is performed on a small amount of data, the impact of such interference on learning can be relatively large. To solve this problem, therefore, we propose a novel 'Character-Independent Transfer-learning' that performs transfer learning independently for each character. In order to confirm the feasibility of the proposed methodology, we performed experiments utilizing the results of pre-training on MSCOCO dataset which is comprised of 120,000 images and about 600,000 general captions. Additionally, according to the advice of an art therapist, about 300 pairs of 'image / expertise captions' were created, and the data was used for the experiments of expertise transplantation. As a result of the experiment, it was confirmed that the caption generated according to the proposed methodology generates captions from the perspective of implanted expertise whereas the caption generated through learning on general data contains a number of contents irrelevant to expertise interpretation. In this paper, we propose a novel approach of specialized image interpretation. To achieve this goal, we present a method to use transfer learning and generate captions specialized in the specific domain. In the future, by applying the proposed methodology to expertise transplant in various fields, we expected that many researches will be actively conducted to solve the problem of lack of expertise data and to improve performance of image captioning.
Various influence-related relationships in Social Network Services (SNS) among users, posts, and user-and-post, can be expressed using links. The current research evaluates the influence of specific users or posts by analyzing the link structure of relevant social network graphs to identify influential users. We applied the concept of mutual interactions proposed for ranking semantic web resources, rather than the voting notion of Page Rank or HITS, to blogosphere, one of the early SNS. Through many experiments with network models, where the performance and validity of each alternative approach can be analyzed, we showed the applicability and strengths of our approach. The weight tuning processes for the links of these network models enabled us to control the experiment errors form the link weight differences and compare the implementation easiness of alternatives. An additional example of how to enter the content scores of commercial or spam posts into the graph-based method is suggested on a small network model as well. This research, as a starting point of the study on identifying influential users in SNS, is distinctive from the previous researches in the following points. First, various influence-related properties that are deemed important but are disregarded, such as scraping, commenting, subscribing to RSS feeds, and trusting friends, can be considered simultaneously. Second, the framework reflects the general phenomenon where objects interacting with more influential objects increase their influence. Third, regarding the extent to which a bloggers causes other bloggers to act after him or her as the most important factor of influence, we treated sequential referencing relationships with a viewpoint from that of PageRank or HITS (Hypertext Induced Topic Selection).
Although the number of patent which is one of the core outputs of technological innovation continues to increase, the number of low-value patents also hugely increased. Therefore, efficient evaluation of patents has become important. Estimation of patent lifespan which represents private value of a patent, has been studied for a long time, but in most cases it relied on a linear model. Even if machine learning methods were used, interpretation or explanation of the relationship between explanatory variables and patent lifespan was insufficient. In this study, patent lifespan (number of renewals) is predicted based on the idea that patent lifespan represents the value of the patent. For the research, 4,033,414 patents applied between 1996 and 2017 and finally granted were collected from USPTO (US Patent and Trademark Office). To predict the patent lifespan, we use variables that can reflect the characteristics of the patent, the patent owner's characteristics, and the inventor's characteristics. We build four different models (Ridge Regression, Random Forest, Feed Forward Neural Network, Gradient Boosting Models) and perform hyperparameter tuning through 5-fold Cross Validation. Then, the performance of the generated models are evaluated, and the relative importance of predictors is also presented. In addition, based on the Gradient Boosting Model which have excellent performance, Accumulated Local Effects Plot is presented to visualize the relationship between predictors and patent lifespan. Finally, we apply Kernal SHAP (SHapley Additive exPlanations) to present the evaluation reason of individual patents, and discuss applicability to the patent evaluation system. This study has academic significance in that it cumulatively contributes to the existing patent life estimation research and supplements the limitations of existing patent life estimation studies based on linearity. It is academically meaningful that this study contributes cumulatively to the existing studies which estimate patent lifespan, and that it supplements the limitations of linear models. Also, it is practically meaningful to suggest a method for deriving the evaluation basis for individual patent value and examine the applicability to patent evaluation systems.
The wall shear stress in the vicinity of end-to end anastomoses under steady flow conditions was measured using a flush-mounted hot-film anemometer(FMHFA) probe. The experimental measurements were in good agreement with numerical results except in flow with low Reynolds numbers. The wall shear stress increased proximal to the anastomosis in flow from the Penrose tubing (simulating an artery) to the PTFE: graft. In flow from the PTFE graft to the Penrose tubing, low wall shear stress was observed distal to the anastomosis. Abnormal distributions of wall shear stress in the vicinity of the anastomosis, resulting from the compliance mismatch between the graft and the host artery, might be an important factor of ANFH formation and the graft failure. The present study suggests a correlation between regions of the low wall shear stress and the development of anastomotic neointimal fibrous hyperplasia(ANPH) in end-to-end anastomoses. 30523 T00401030523 ^x Air pressure decay(APD) rate and ultrafiltration rate(UFR) tests were performed on new and saline rinsed dialyzers as well as those roused in patients several times. C-DAK 4000 (Cordis Dow) and CF IS-11 (Baxter Travenol) reused dialyzers obtained from the dialysis clinic were used in the present study. The new dialyzers exhibited a relatively flat APD, whereas saline rinsed and reused dialyzers showed considerable amount of decay. C-DAH dialyzers had a larger APD(11.70
The wall shear stress in the vicinity of end-to end anastomoses under steady flow conditions was measured using a flush-mounted hot-film anemometer(FMHFA) probe. The experimental measurements were in good agreement with numerical results except in flow with low Reynolds numbers. The wall shear stress increased proximal to the anastomosis in flow from the Penrose tubing (simulating an artery) to the PTFE: graft. In flow from the PTFE graft to the Penrose tubing, low wall shear stress was observed distal to the anastomosis. Abnormal distributions of wall shear stress in the vicinity of the anastomosis, resulting from the compliance mismatch between the graft and the host artery, might be an important factor of ANFH formation and the graft failure. The present study suggests a correlation between regions of the low wall shear stress and the development of anastomotic neointimal fibrous hyperplasia(ANPH) in end-to-end anastomoses. 30523 T00401030523 ^x Air pressure decay(APD) rate and ultrafiltration rate(UFR) tests were performed on new and saline rinsed dialyzers as well as those roused in patients several times. C-DAK 4000 (Cordis Dow) and CF IS-11 (Baxter Travenol) reused dialyzers obtained from the dialysis clinic were used in the present study. The new dialyzers exhibited a relatively flat APD, whereas saline rinsed and reused dialyzers showed considerable amount of decay. C-DAH dialyzers had a larger APD(11.70
Purpose : This study has already started commercial Gated RapidArc automation equipment which was not previously in the Gated radiation therapy can be performed simultaneously with the VMAT Gated RapidArc radiation therapy to the accuracy of the analysis to evaluate the usability, Amplitude mode applied to the patient. Materials and Methods : The analysis of the distribution of radiation dose equivalent quality solid water phantom and GafChromic film was used Film QA film analysis program using the Gamma factor (3%, 3 mm). Three-dimensional dose distribution in order to check the accuracy of Matrixx dosimetry equipment and Compass was used for dose analysis program. Periodic breathing synchronized with solid phantom signals Phantom 4D Phantom and Varian RPM was created by breathing synchronized system, free breathing and breath holding at each of the dose distribution was analyzed. In order to apply to four patients from February 2013 to August 2013 with liver cancer targets enough to get a picture of 4DCT respiratory cycle and then patients are pratice to meet patient's breathing cycle phase mode using the patient eye goggles to see the pattern of the respiratory cycle to be able to follow exactly in a while 4DCT images were acquired. Gated RapidArc treatment Amplitude mode in order to create the breathing cycle breathing performed three times, and then at intervals of 40% to 60% 5-6 seconds and breathing exercises that can not stand (Fig. 5), 40% While they are treated 60% in the interval Beam On hold your breath when you press the button in a way that was treated with semi-automatic. Results : Non-respiratory and respiratory rotational intensity modulated radiation therapy technique absolute calculation dose of using computerized treatment plan were shown a difference of less than 1%, the difference between treatment technique was also less than 1%. Gamma (3%, 3 mm) and showed 99% agreement, each organ-specific dose difference were generally greater than 95% agreement. The rotational intensity modulated radiation therapy, respiratory synchronized to the respiratory cycle created Amplitude mode and the actual patient's breathing cycle could be seen that a good agreement. Conclusion : When you are treated Non-respiratory and respiratory method between volumetric intensity modulated radiation therapy rotation of the absolute dose and dose distribution showed a very good agreement. This breathing technique tuning volumetric intensity modulated radiation therapy using a rotary moving along the thoracic or abdominal breathing can be applied to the treatment of tumors is considered. The actual treatment of patients through the goggles of the respiratory cycle to create Amplitude mode Gated RapidArc treatment equipment that does not automatically apply to the results about 5-6 seconds stopped breathing in breathing synchronized rotary volumetric intensity modulated radiation therapy facilitate could see complement.