A Data Organization Method for LSTM and Transformer When Predicting Chinese Banking Stock Prices

The accurate prediction of stock prices is not an easy task. The long short-term memory (LSTM) neural network and the transformer are good machine learning models for times series forecasting. In this paper, we use LSTM and transformer to predict prices of banking stocks in China’s A-share market. It is shown that organizing the input data can help get accurate outcomes of the models. In this paper, we first introduce some basic knowledge about LSTM and present prediction results using a standard LSTM model. Then, we show how to organize the input data during the training period and give the comparison results for not only LSTM but also the transformer model. The numerical results show that the prediction results of LSTM and transformer can be improved after the input data are organized when training.

Product design for people with Rheumatoid Arthritis and its relationship with stakeholder’s expectations

Approximately 6.2% of the Brazilian population has some disability, from which 1.3% declare to have some physical disability. Rheumatoid Arthritis (RA) is a systemic and autoimmune disease that affects about 0.4% to 1.3% of the world population, being more common in women. In advanced stages, the disease can cause severe impairment in small joints, limiting individuals' movements and impairing their independence and autonomy. Assistive Technology (AT) aims to promote actions that enable People with Disabilities or Reduced Mobility to execute their daily activities in the most normal and independent manner possible. However, studies show high rates of abandonment of AT devices for reasons generally related to the effectiveness of the product itself. Thus, this study aims to carry out data collection, organization, and analysis to understand the expectations of stakeholders about the development of new AT devices for individuals with RA. For this, the technical research procedures were segmented as follows: Phase 1 – Theoretical Foundation; Phase 2 – Selection of Target Audience and Sampling; Phase 3 – Data Survey; and Phase 4 – Data Organization and Analysis. Through the procedures adopted, it was possible to verify that there are significant differences between the expectations of stakeholders concerning the development of new AT products for individuals with RA, which can directly imply user satisfaction, consequently generating the abandonment of these devices if they do not meet the expectations.

The regenerating protein 3A: a crucial molecular with dual roles in cancer

Introduction REG3A, a member of the third subclass of the Reg family, has been found in a variety of tissues but is not detected in immune cells. In the past decade, it has been determined that REG3A expression is regulated by injury, infection, inflammatory stimuli, and pro-cytokines via different signaling pathways, and it acts as a tissue-repair, bactericidal, and anti-inflammatory molecule in human diseases. Recently, the role of REG3A in cancer has received increasing attention. The present article aims to investigate the structure, expression, regulation, function of REG3A, and to highlight the potential role of REG3A in tumors. Methods A detailed literature search and data organization were conducted to find information about the role of REG3A in variety of physiological functions and tumors. Results Contradictory roles of REG3A have been reported in different tumor models. Some studies have demonstrated that high expression of REG3A in cancers can be oncogenic. Other studies have shown decreased REG3A expression in cancer cells as well as suppressed tumor growth. Conclusions Taken together, better understanding of REG3A may lead to new insights that make it a potentially useful target for cancer therapy.

Co-operative categorization in civil engineering

In the paper we indicate the way to allow the coexistence of multiple structures that are mutually supportive, but not mutually exclusive. It allows less organized users to classify entities according to their suitable structure, whilst leaving the possibility for specialised users to classify these entities into related field classification systems or domain ontologies, in mutually beneficial cooperation. Replacing the reductionist enumerative structure with multiplicative categorization should bring organizational and terminological flexibility that will allow future updates without disrupting the existing categorization. It allows one to use parallel hierarchical branches of categories according to aspects and classify the entity into several branches at the same time. It should allow users the freedom to create their own structures without disrupting the structures of other user groups. Such structures could coexist happily side by side. The aim is to indicate the model of an independent data organization system in building design in a multilingual environment, connected through diverse CAD, GIS, BIM, CAFM applications, ECM, CDE, emails, databases and file systems.

Data organization issues in civil engineering

In today's organization of data in civil engineering there remains ambiguity, diversity of terminology and consequent lack of clarity. The daily reality is to work with hybrid data sources in a variety of software programs. In CAD programs it is possible to import and convert elements from GIS, but they will be classified only into layers. Or, attach external references from CAD files in BIM, but there still remains the problem of finding your way around the confusing list of layers, and their abbreviations. Also broadly used filesystems as place for store and organise data on disk into files (entities) and hierarchical structure of folders (categories) is limited by reductionist categorization approach. Another problem shows with absence of multilingual approach. The limits are in the methods of data organization, and so only partially portable. Therefore, if there is a common denominator to all problems, it is the absence of a unifying organizational environment. In current computer applications used in civil engineering, methods of data classification are implemented that do not allow multiplicative or faceted categorization and thus inevitably create an antagonistic environment with any previous or future classification system. Unfortunately, even if a software tool implements a modern multiplicative categorization or domain ontology, it will not significantly solve the overall situation, because it will still not be possible to use such a method of categorization in other applications. Therefore, if there should exist a solution that would significantly facilitate cooperation and orientation in the data, it must inevitably work independently, across diverse software applications, with the ability to link the functions of the application with such a separate categorization system.

Efficient Visualization of Large-Scale Oblique Photogrammetry Models in Unreal Engine

Oblique photogrammetry models are indispensable for implementing digital twins of cities. Geographic information system researchers have proposed plenty of methods to load and visualize these city-scaled scenes. However, when the area viewed changes quickly in real-time rendering, current methods still require excessive GPU calculation and memory occupation. In this study, we propose a data organization method in which we merged all quadtrees and used a binary encoding method to encode nodes in a merged tree so that the parent–child relationship between the tree nodes could be calculated using rapid binary operations. After that, we developed a strategy to cancel the loading of redundant nodes based on the parent–child relationship, which helped to reduce the hard disk loading time and the amount of memory occupied in visualization. Moreover, we introduced a parameter to measure the area of the triangle mesh per pixel to achieve unified data scheduling under different production standards. We implemented our method based on Unreal Engine (UE), and three experiments were designed to illustrate the advantages of our methods in index acceleration, frame time, and memory reduction. The results show that our methods can significantly improve visualization fluency and reduce memory usage.