• Journal of the Society of Naval Architects of Korea
• /
• v.58 no.5
• /
• pp.281-293
• /
• 2021

Oil and steel prices, which are major pricescosts in the shipbuilding industry, were predicted. Firstly, the error of the moving average line (N=3-5) was examined, and in all three error analyses, the moving average line (N=3) was small. Secondly, in the linear prediction of data through existing theory, oil prices rise slightly, and steel prices rise sharply, but in reality, linear prediction using existing data was not satisfactory. Thirdly, we identified the limitations of linear prediction methods and confirmed that oil and steel price prediction was somewhat similar to actual moving average line prediction methods. Due to the high volatility of major price flows, large errors were inevitable in the forecast section. Through the time series analysis method at the end of this paper, we were able to achieve not bad results in all analysis items relative to artificial intelligence (Prophet). Predictive data through predictive analysis using eight predictive models are expected to serve as a good research foundation for developing unique tools or establishing evaluation systems in the future. This study compares the basic settings of artificial intelligence programs with the results of core price prediction in the shipbuilding industry through time series prediction theory, and further studies the various hyper-parameters and event effects of Prophet in the future, leaving room for improvement of predictability.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.13 no.1
• /
• pp.101-107
• /
• 2013

Is generated in real time in the real world, a large amount of time series data from a wide range of business areas. But it is not easy to determine the optimal model for the description and understanding of the time series data is represented as a dynamic feature. In this study, through the HMM suitable for estimating the short and long-term forecasting model of time-series data to estimate a model that can explain the characteristics of these time series data, it was estimated to predict future patterns of movement. The actual stock market through various materials, information criterion and optimal model estimation for the length of the most efficient data was found to accurately estimate the state of the model. Similar movement patterns predictive than the long-term prediction is more similar to the short-term prediction of the experimental result were found to be.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.1
• /
• pp.1-14
• /
• 2024

Time-series forecasting is extensively used in the actual world. Recent research has shown that Transformers with a self-attention mechanism at their core exhibit better performance when dealing with such problems. However, most of the existing Transformer models used for time series prediction use the traditional encoder-decoder architecture, which is complex and leads to low model processing efficiency, thus limiting the ability to mine deep time dependencies by increasing model depth. Secondly, the secondary computational complexity of the self-attention mechanism also increases computational overhead and reduces processing efficiency. To address these issues, the paper designs an efficient multi-layer attention-based time-series forecasting model. This model has the following characteristics: (i) It abandons the traditional encoder-decoder based Transformer architecture and constructs a time series prediction model based on multi-layer attention mechanism, improving the model's ability to mine deep time dependencies. (ii) A cross attention module based on cross attention mechanism was designed to enhance information exchange between historical and predictive sequences. (iii) Applying a recently proposed sparse attention mechanism to our model reduces computational overhead and improves processing efficiency. Experiments on multiple datasets have shown that our model can significantly increase the performance of current advanced Transformer methods in time series forecasting, including LogTrans, Reformer, and Informer.

• Journal of the Korean Data and Information Science Society
• /
• v.18 no.1
• /
• pp.17-29
• /
• 2007

The major purpose of this article is the comparison of estimation method with Newton-Raphson, Kalman-filter, and prediction method with Kalman prediction. Conditional expectation in space time bilinear(STBL) model, which is a very powerful and parsimonious nonlinear time-series model for the space time series data can be viewed as a set of time series collected simultaneously at a number of spatial locations and time points, and which have appeared in a important applications areas: geography, geology, natural resources, ecology, epidemiology, etc.

• Proceedings of the KIEE Conference
• /
• 2004.07d
• /
• pp.2209-2211
• /
• 2004

Unlike the wavelet neural network, since a mother wavelet layer of the self-recurrent wavelet neural network (SRWNN) is composed of self-feedback neurons, it has the ability to store past information of the wavelet. Therefore we propose the prediction method for the nonlinear chaotic time series model using a SRWNN. The SRWNN model is learned for the modeling of a function such that the inputs arc known values of the time series and the output is the value in the future. The parameters of the network are tuned to minimize the difference between the nonlinear mapping of the chaotic time series and the output of SRWNN using the gradient-descent method for the adaptive backpropagation algorithm. Through the computer simulations, we demonstrate the feasibility and the effectiveness of our method for the prediction of the logistic map and the Mackey-Glass delay-differential equation as a nonlinear chaotic time series.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.12
• /
• pp.579-586
• /
• 2021

Time series forecasting refers to predicting future time information based on past time information. Accurately predicting future information is crucial because it is used for establishing strategies or making policy decisions in various fields. Recently, a transformer model has been mainly studied for a time series prediction model. However, the existing transformer model has a limitation in that it has an auto-regressive structure in which the output result is input again when the prediction sequence is output. This limitation causes a problem in that accuracy is lowered when predicting a distant time point. This paper proposes a sequential decoding model focusing on the style transformation technique to handle these problems and make more precise time series forecasting. The proposed model has a structure in which the contents of past data are extracted from the transformer-encoder and reflected in the style-based decoder to generate the predictive sequence. Unlike the decoder structure of the conventional auto-regressive transformer, this structure has the advantage of being able to more accurately predict information from a distant view because the prediction sequence is output all at once. As a result of conducting a prediction experiment with various time series datasets with different data characteristics, it was shown that the model presented in this paper has better prediction accuracy than other existing time series prediction models.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.11
• /
• pp.1471-1476
• /
• 2021

Flood damage can cause floods or tsunamis, which can result in enormous loss of life and property. In this regard, damage can be reduced by making a quick evacuation decision through flood prediction, and many studies are underway in this field to predict floods using time series data. In this paper, we propose a CNN-based time series prediction model. A CNN-based water level prediction model was implemented using the river level and precipitation, and the performance was confirmed by comparing it with the LSTM and GRU models, which are often used for time series prediction. In addition, by checking the performance difference according to the size of the input data, it was possible to find the points to be supplemented, and it was confirmed that better performance than LSTM and GRU could be obtained. Through this, it is thought that it can be utilized as an initial study for flood prediction.

• The Korean Journal of Applied Statistics
• /
• v.32 no.6
• /
• pp.851-865
• /
• 2019

Discovering various features from one time series is complicated. In this paper, we introduce a multi aggregation prediction algorithm (MAPA) that uses the concepts of temporal aggregation and combining forecasts to find multiple patterns from one time series and increase forecasting accuracy. Temporal aggregation produces multiple time series and each series has separate properties. We use exponential smoothing methods in the next step to extract various features of time series components in order to forecast time series components for each series. In the final step, we blend predictions of the same kind of components and forecast the target series by the summation of blended predictions. As an empirical example, we forecast traffic accident counts using MAPA and observe that MAPA performance is superior to conventional methods.

• The Korean Journal of Applied Statistics
• /
• v.26 no.6
• /
• pp.1019-1032
• /
• 2013

The accurate prediction of future mortality is an important issue due to recent rapid increases in life expectancy. An accurate estimation and prediction of mortality is important to future welfare policies. The optimal selection of a mortality model is important to estimate and predict mortality; however, the period of time series data used is also an important issue. It is essential to understand that the time series data for mortality is short in Korea and the data before 1982 is incomplete. This paper divides the time series of Korean mortality into two sets to compare the parameter estimates of the LC model and LC model with a cohort effect by the period of data used. A modeling and prediction of the mortality index and cohort effect index as well as the evaluation of future life expectancy is conducted. Finally, some suggestions are proposed for the future prediction of mortality.

• The Journal of Information Systems
• /
• v.29 no.1
• /
• pp.241-265
• /
• 2020

Purpose Various machine learning techniques are used to implement for predicting corporate credit. However, previous research doesn't utilize time series input features and has a limited prediction timing. Furthermore, in the case of corporate bond credit rating forecast, corporate sample is limited because only large companies are selected for corporate bond credit rating. To address limitations of prior research, this study attempts to implement a predictive model with more sample companies, which can adjust the forecasting point at the present time by using the credit score information and corporate information in time series. Design/methodology/approach To implement this forecasting model, this study uses the sample of 2,191 companies with KIS credit scores for 18 years from 2000 to 2017. For improving the performance of the predictive model, various financial and non-financial features are applied as input variables in a time series through a sliding window technique. In addition, this research also tests various machine learning techniques that were traditionally used to increase the validity of analysis results, and the deep learning technique that is being actively researched of late. Findings RNN-based stateful LSTM model shows good performance in credit rating prediction. By extending the forecasting time point, we find how the performance of the predictive model changes over time and evaluate the feature groups in the short and long terms. In comparison with other studies, the results of 5 classification prediction through label reclassification show good performance relatively. In addition, about 90% accuracy is found in the bad credit forecasts.