KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Settings

Subject: search.filters.subject.Multilayer perceptron

Search Results

Now showing 1 - 3 of 3
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Multilayer perceptron-based detection for a coded two-dimensional magnetic recording
    (2023-02-01)
    Buajong, Chaiwat
    ;
    Warisarn, Chanon
    Recently, magnetic recording technology has been on the verge of reaching its recording density limit. Two-dimensional (2-D) magnetic recording (TDMR) is expected to be a candidate for next-generation magnetic recording because it provides powerful 2-D signal processing tools that are seamlessly compatible with the current technology. However, the degradation in signal quality due to 2-D interference when increasing the recording density far beyond the limit is inevitable even advanced signal processing tools cannot thoroughly handle such interference. In this work, we introduce two architectures of the multilayer perceptron (MLP)-based detection capable of producing soft information in the form of a log-likelihood ratio (LLR) for a coded TDMR system. This also enables turbo decoding capability to take place in the system, which further enhances the system performance. Such an architecture has two structures explicitly designed to execute before and during turbo decoding with a low-density parity-check code. In the first architecture, we employ triple MLP-based detection to process three readback sequences individually. For another architecture, it is designed to collectively process three readback sequences using just single MLP-based detection. The results show that both architectures of proposed detections outperform the systems using the conventional detection based on the Viterbi algorithm and single MLP-based detection achieves slightly better performance than triple MLP-based detection despite having much less complexity.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Neural Networks Input Techniques to Maintain a Small Skew Angle in Bit-Patterned Magnetic Recording with a V-Shaped Read-Head Array
    (2023-01-01)
    Fatika, Kirana Alif
    ;
    Koonkarnkhai, Santi
    ;
    Kovintavewat, Piya
    ;
    Warisarn, Chanon
    The demand for enormous storage devices has kept increasing, leading to the development of various advanced technologies with a vast storage capacity. Extensive numbers of related research studies have been aiming at optimizing code design and algorithms analytically; however, enacting them on practical devices has been scarce. Achieving this demand might bring some obstacles called two-dimensional interference and skew angle (SA). To meet the challenge of the obstacle, we propose a SA detection method for bit-patterned magnetic recording systems by computing a specific target by three readback sequences before estimating the SA value and detecting the SA amount happening in the system using an application of neural network namely multilayer perceptron. An error correction code, low-density parity-check, is applied, and its decoder outputs a log-likelihood ratio whose probability density distribution is examined. The simulation results show that the sliding window technique can significantly provide a better bit error rate performance.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    A deep learning model for predicting buy and sell recommendations in stock exchange of Thailand using long short-term memory
    (2019-02-01)
    Sanboon, Thaloengpattarakoon
    ;
    Keatruangkamala, Kamol
    ;
    Jaiyen, Saichon
    Nowadays, the stock price prediction has been one of the most challenging problem to the AI research community. Most prediction techniques concentrate on forecasting the future prices of stocks based on conventional Machine learning techniques. However, these techniques cannot capture long term dependencies in stock price data. Therefore, they cannot consider the relation between the current predicted data and the previous data in stock data. This research adopts deep learning techniques for predicting buy and sell recommendations in Stock Exchange of Thailand using Long Short-Term Memory. The proposed model can capture long term dependencies in stock price data in order to enhance the prediction accuracy. The accuracy of the proposed model is evaluated on five Stock Exchange of Thailand (SET) stocks, between 5 January 2015 and 29 December 2017, and compared the results with support vector Machine, multilayer perceptron, decision tree, random forest, logistic regression and k-nearest neighbors. The experimental results signify that the proposed model can outperform all comparative models.