Narrow Results

Temperature is an essential quality index in storage. Prediction of temperature can help the grain storage industry to apply the appropriate operations such as ventilation or drying to improve the quality of grain and extend the suitable storage time. Traditional machine learning methods usually cannot accurately predict the temperature data of the grain considering the complexity of environmental factors and grain warehouse conditions. To make better use of the temporal data such as temperature/humidity information of grain itself and its environment, this paper proposes a gated recurrent unit (GRU)-based algorithm to predict the change of the data. The grain warehouse environmental data are collected by multi-functional sensors inside a grain depot, including temperature, humidity, wind speed, air pressure, etc. Some of these data features such as rain or snow days are sparse data features. Excessive sparse features can affect the training accuracy of the model. At the same time, due to sensor aging or extreme weather conditions, the data collected may not be accurate, and the data contain noise, which also has a significant impact on the training of the model. To improve the performance of the proposed GRU framework, multivariate linear regression is used for feature generation to optimize the volatility of weather data, strengthen and construct the characteristics of datasets, and wavelet filtering is used to denoise the corresponding features. This paper focuses on the data sparse and noise problem and applies the MLR and wavelet filtering to improve the GRU prediction framework for grain warehouse temporal data. According to our experiment, the temperature prediction results based on the GRU deep fusion model have better improvement in prediction accuracy and time than the existing neural network algorithms such as long–short-term memory (LSTM), GRU, and transformer.

Highly reliable software is becoming an essential ingredient in many systems. However, assuring reliability often entails time-consuming costly development processes. One cost-effective strategy is to target reliability-enhancement activities to those modules that are likely to have the most problems. Software quality prediction models can predict the number of faults expected in each module early enough for reliability enhancement to be effective.

This paper introduces a case-based reasoning technique for the prediction of software quality factors. Case-based reasoning is a technique that seeks to answer new problems by identifying similar "cases" from the past. A case-based reasoning system can function as a software quality prediction model. To our knowledge, this study is the first to use case-based reasoning systems for predicting quantitative measures of software quality.

A case study applied case-based reasoning to software quality modeling of a family of full-scale industrial software systems. The case-based reasoning system's accuracy was much better than a corresponding multiple linear regression model in predicting the number of design faults. When predicting faults in code, its accuracy was significantly better than a corresponding multiple linear regression model for two of three test data sets and statistically equivalent for the third.

Cross-project defect prediction (CPDP) has recently become very popular in the field of software defect prediction. It was generally treated as a binary classification problem or a regression problem in most of previous studies. However, these existing CPDP methods may be not suitable for those software projects that have limited manpower and budget. To address the issue of priority estimation for buggy software entities, in this paper CPDP is formulated as a ranking problem. Inspired by the idea of the pointwise approach to learning to rank, we propose a ranking-oriented CPDP approach called ROCPDP. A case study conducted on the datasets collected from AEEEM and PROMISE shows that ROCPDP outperforms the eight baseline methods in two CPDP scenarios, namely one-to-one and many-to-one. Besides, in the many-to-one scenario ROCPDP is, by and large, comparable to the best baseline method performed in a specific within-project defect prediction scenario.

Long-term care for the elderly has become one of the prominent social problems globally when the ratios of persons whose ages over 65 steadily increase in almost all countries. One of the solution approaches that could be adapted is called long-term care insurance provided by insurance companies. However, companies need to classify care status types based on price or to provide supports utilizing its organizational structures such as departmental communication, business selection, and market segmentation since long-term care consists of many factors. The motivation of this research aims at filling the gap since there exists no comprehensive research concerning these factors that have impacts on the long-term care status for the elderly. To determine those factors, machine learning (ML) algorithms such as multiple linear regression, random forest, and the XGBoost are selected to be employed. Then, those factors and their important variables are utilized to predict insurance pricing. The 2018 Chinese (CHARLS) data set is used to determine factors that have key impacts on long-term care status in the elderly. Finally, all models are combined as a comprehensive model to generate better prediction accuracies innovatively. The results show that the three ML models can provide relatively consistent important measures of risk factors in determining the nursing status of the elderly. On the other hand, the prediction accuracy of the random forest and the XGBoost was improved by 0.6% and 1%, respectively, when compared to multiple linear regression. Besides, the results show that when the ratios of 2.6, 3.7, 3.7 are assigned to the results of the three models, the prediction accuracy of the comprehensive model is higher in the test set than that of the multiple linear regression, which contributes 1.92% more. The main innovation of this research is to construct a comprehensive model, a weighted combination of three models, with better prediction accuracy. Eventually, the long-term care insurance business can utilize the comprehensive model to classify the long-term care status of the elderly.

Lubricating additives can improve the lubricant performance of base oil in reducing friction and wear and minimizing loss of energy. It is of great significance to study the relationship between chemical structures and lubrication properties of lubricant additives. This paper reports a quantitative structure–property relationship (QSPR) model of the maximum nonseizure loads ($P_{\mathrm{\text{B}}}$) of 79 lubricant additives by applying artificial neural network (ANN) based on the algorithm of backward propagation of errors. Six molecular descriptors appearing in the multiple linear regression (MLR) model were used as vectors to develop the ANN model. The optimal condition of ANN with network structure of [6-4-1] was obtained by adjusting various parameters by trial-and-error. The root-mean-square (rms) errors from ANN model are $52.9N$ ($R=0.965$) for the training set and $61.4N$ ($R=0.940$) for the test set, which are superior to the MLR results of $98.5N$ ($R=0.873$) for the training set and $95.6N$ ($R=0.866$) for the test set. Compared to the existing model for $P_{\mathrm{\text{B}}}$, our model has better statistical quality. The results indicate that our ANN model can be applied to predict the $P_{\mathrm{\text{B}}}$ values for lubricant additives.

It is difficult to find rules in the huge stock system, but we can enter from a small entrance, explore the inextricable links between the market return and some time series variables, and predict the overall return based on these variables. This paper first explains the relationship between the following economic and financial time series variables and market return, then explain the reasons for choosing them: CPI, momentum, dividend rate, market book ratio, P/E ratio, a turnover rate of trading volume calculation, market-capitalization-weighted idiosyncratic volatility, logarithm of stock market liquidity index, Amihud illiquidity index, skew; secondly, this paper uses Python to make descriptive statistical analysis of these variable data; Thirdly, it plots variables and yields to compare trends; Then, it constructs a neural network model by multiple regression to select significantly correlated variables; Finally, the BP neural network is used to predict the return of China’s stock market.

Investor confidence can influence the decision-making behavior of investors themselves and others to a certain extent. Based on the data on investor confidence and commercial credit financing of A-share listed enterprises in China from 2010 to 2021, this paper uses a multiple linear regression method to establish a model and observes the multiple linear regression relationship with the help of stata16 metrology software. The results show that there is a significant positive correlation between investor confidence and the commercial credit financing ability of enterprises. This relationship only exists in non-state-owned enterprises, and the lower the concern of enterprises in research reports, the stronger the promotion effect of investor confidence on commercial credit financing.

As the integration development of the Yangtze River Delta area is defined as a national strategy, both the “integration” and “high level” have become the common development goals of the three provinces and one municipality within the area. Therefore, how to promote high-quality agricultural integration development with the internal resource flow within the Yangtze River Delta so as to solve the problems of large differences and imbalances in the agricultural development in this area is of great importance. Based on the policy orientation for agricultural development, the fiscal expenditure related to agriculture, the total agricultural machinery power, and the cultivated land area are selected as the three independent indicators to analyze the impact they would make on the total agricultural output. By using a multiple linear regression model, we have found that both fiscal expenditures related to agriculture and the cultivated land area will make a significantly positive impact on the total output, while the machinery power will help to improve the output value to some extent.

By collecting financial big data from the Wind database, this paper takes the panel data of listed companies on the SME Board and GEM Board of Shenzhen Market from 2011 to 2020 as samples, uses the multiple linear regression method to construct a model for empirical analysis, and observes the relationship between multiple linear regression to explore the impact of R&D expenditure on future earnings. The results show that: (1) R&D expenditure positively correlates with future earnings. R&D expenditure has a positive impact on future income level. (2) The correlation between R&D expenditure and future earnings has decreased significantly over time. The impact has weakened with time, which means that the R&D income of listed companies in Shenzhen has shown a downward trend in recent years.

The following sections are included:

• INTRODUCTION

• THE MODEL AND ITS ASSUMPTIONS
  • The Multiple Regression Model
  • The Regression Plane for Two Explanatory Variables
  • Assumptions for the Multiple Regression Model

• ESTIMATING MULTIPLE REGRESSION PARAMETERS
  • Annual Salary, Years of Education, and Years of Work Experience

• THE RESIDUAL STANDARD ERROR AND THE COEFFICIENT OF DETERMINATION
  • The Residual Standard Error
  • The Coefficient of Determination

• TESTS ON SETS AND INDIVIDUAL REGRESSION COEFFICIENTS
  • Test on Sets of Regression Coefficients
  • Hypothesis Tests for Individual Regression Coefficients

• CONFIDENCE INTERVAL FOR THE MEAN RESPONSE AND PREDICTION INTERVAL FOR THE INDIVIDUAL RESPONSE
  • Point Estimates of the Mean and the Individual Responses
  • Interval Estimates of Forecasts

• BUSINESS AND ECONOMIC APPLICATIONS
  • Overall Job-Worth of Performance for Certain Army Jobs
  • The Relationship Between Individual Stock Rates of Return, Payout Ratio, and Market Rates of Return
  • Analyzing the Determination of Price per Share
  • Multiple Regression Approach to Evaluating Real Estate Property
  • Multiple Regression Approach to Doing Cost Analysis

• USING COMPUTER PROGRAMS TO DO MULTIPLE REGRESSION ANALYSES
  • SAS Program for Multiple Regression Analysis
  • MINITAB Program for Multiple Regression Prediction
  • Stepwise Regression Analysis

• Summary

• Appendix 15A Derivation of the Sampling Variance of the Least-Squares Slope Estimations

• Appendix 15B Derivation of Equation 15.30

• Questions and Problems

Most power estimation methods of the traditional coarse-grained reconfigurable processors use simulation method, although the accuracy is higher, but the efficiency is very low. According to the characteristics of coarse grained reconfigurable computing processor, this paper proposes a new method of power estimation based on the macro power model. This method builds a multiple linear equation which is based on the theory of multiple linear regression, and estimates the total power of the processor through the equation. Macro power model consists of variable and regression coefficients. Data word width, working frequency and intensive calculation amount of application algorithm as parameter variable are chosen. They have important impact on power. Regression coefficients express the contribution degree of variable to power. It could be solved by low level gate simulation under different input vectors. The macro power model of reconfigurable pipelined processor is constructed by using the architecture of coarse grained reconfigurable pipelined processor as an example. Matrix multiplication is used to test the different combinations of macro power model variables, and the average error between the estimated and the actual analog values is 14.12%, which speed is hundred times faster than the traditional simulation method. The experimental results demonstrate that the macro power model is an efficient and accurate power estimation method.