Table of Content

  

20 July 2021
Volume 56 Issue 7

Personalized hotel recommendation algorithm based on user implicit data

SHI Da, YU Miao-chuan, LI Meng-qi

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  1-10.  doi:10.6040/j.issn.1671-9352.0.2021.151

Abstract ( 1023 )   PDF (1528KB) ( 675 )   Save

References | Related Articles | Metrics

Based on collaborative filtering of items, this paper mines and models the implicit feedback data, designs the implicit feedback preference scoring rules, and gives a new definition to calculate the hotel similarity formula. At the same time, considering the basic characteristics of users will also have an impact on users personalized needs and the limitations of a single algorithm, this paper further introduces the XGBoost model, and filters the improved recommendation results with XGBoost training, so as to obtain a better personalized hotel recommendation system. This paper adopts real desensitization data to prove that the recommendation effect of building personalized hotel recommendation system based on cascade model is more accurate, which has a strong reference value for the personalized service of online hotel platform.

Research on optimization of transportation routes in large logistics parks considering carrying capacity: taking a logistics park in Linyi as an example

ZOU Nan, XIE Lei, JIN Zong-kai, CAI Xing-mao, SHAO Chen

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  11-20.  doi:10.6040/j.issn.1671-9352.0.2021.150

Abstract ( 1228 )   PDF (1907KB) ( 379 )   Save

References | Related Articles | Metrics

In order to solve the problem of traffic congestion in the park, a multi-objective optimization model is constructed to reduce transportation costs and improve traffic congestion. The results show that optimization methods can effectively reduce transportation costs and improve traffic congestion. Based on the analysis of the model and the results, it provides suggestions for the construction and improvement measures of the logistics park.

Product pricing and operating decision of manufacturers under the sharing

SONG Xiao-tong, ZHANG Gui-tao, WANG Guang-qin

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  21-31.  doi:10.6040/j.issn.1671-9352.0.2020.585

Abstract ( 982 )   PDF (1319KB) ( 402 )   Save

References | Related Articles | Metrics

Based on the B2C sharing model, the manufacturer decides whether to sell or rent the product. Establish the duopoly game model, analyze the influence of different channel management strategies on manufacturers prices and profits, and get the conditions to meet the requirements of each equilibrium strategy. The results show that, except for the pure rental strategy combination, the purchase price is positively correlated with the purchase preference, while the lease price is on the contrary. The rental frequency does not affect the pricing. The manufacturers channel sales strategy is influenced by the lease times, market transfer rate and market preference. With the increase of the lease times, the equilibrium range of pure sales strategy combination is first reduced, then increased and finally reduced, and the equilibrium range of pure lease strategy combination keeps increasing.

Complex dynamic analysis of the duopoly game under management delegation

ZHU Yan-lan, ZHOU Wei, CHU Tong, LI Wen-na

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  32-45.  doi:10.6040/j.issn.1671-9352.0.2020.293

Abstract ( 931 )   PDF (14369KB) ( 453 )   Save

References | Related Articles | Metrics

On the basis of the exponential inverse demand function and considering that the enterprise managers take the maximization of their own interests as the objective function, the Cournot dynamic oligopoly game model which is more in line with the actual market environment is established. Through calculation, the system has four equilibrium points, and their local stabilities are analyzed using nonlinear dynamics theory and Jury criterion. The influence of parameters such as the speed of adjusting, weight coefficient and marginal cost on the stability of the system and the evolution process of the system under the change of parameters are analyzed by one-dimensional bifurcation diagrams and two-dimensional bifurcation diagrams. Three kinds of intermittent chaos phenomena are found, and it is concluded that this is a way of self-regulation of the system. In addition, multistability of the system is discussed through the basin of attractive, and it is concluded that the selection of initial conditions will affect the final state of system.

Bayesian inference of Poisson distribution based on left truncated and right censored data

HU Jiang-shan, SUI Yun-yun, FU Yun-peng

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  46-52.  doi:10.6040/j.issn.1671-9352.0.2020.390

Abstract ( 907 )   PDF (406KB) ( 461 )   Save

References | Related Articles | Metrics

The Bayesian inference of Poisson distribution in left truncated and right censored data is studied. The maximum likelihood estimation method and Bayesian estimation are respectively used to estimate the unknown parameters. Meanwhile the corresponding confidence intervals are given, and stochastic simulations are carried out. It is found that in the case of small sample, the accuracy of Bayesian estimator is higher than that of maximum likelihood estimator, while in the case of large sample, the accuracy of these two estimators is not much different. In terms of confidence interval, the confidence interval of the highest posterior probability density constructed is indeed more effective than the traditional confidence interval, no matter it is a small sample or a large sample.

Research on characteristics of air quality change in Zhejiang Province——based on functional data analysis

WU Qi-ran, ZHOU Li-kai, SUN Jin-jin, WANG Nian-ge, YU Qun-fang

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  53-64.  doi:10.6040/j.issn.1671-9352.0.2021.232

Abstract ( 1130 )   PDF (16831KB) ( 370 )   Save

References | Related Articles | Metrics

In order to explore the characteristics of air quality changes in Zhejiang Province, this paper selects the monthly air quality index data and the monthly data of the concentration of six major air pollutants in Zhejiang Province from 2014 to 2019 for research. First, this article fits the air quality index data and the six air pollutant concentration data into a function curve. Then, in order to discover the dynamic change characteristics of the air quality index and the concentration of six air pollutants in Zhejiang Province, this paper performs K-means clustering on the principal component basis coefficients of the air quality index curve family, and analyzes the six air pollutant concentration curve families perform multivariate Funclust clustering. Finally, we used ArcGIS to draw the spatial distribution map of the clustering results in Zhejiang Province and explored the spatial distribution characteristics of the air quality in Zhejiang Province. The results show that the characteristics of air quality changes in Zhejiang Province can be divided into 4 types of regions spatially, showing a trend of gradual deterioration from south to north; in terms of time, the air quality index of Zhejiang Province has been steadily decreasing year by year, and the concentration of SO₂ has dropped significantly. The concentrations of the other five pollutants have obvious similarities and significant seasonal characteristics.

Generalized characteristic polynomial of subdivision Q-neighbourhood corona graphs and its application

YANG Ying, LI Mu-chun, ZHANG You

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  65-72.  doi:10.6040/j.issn.1671-9352.0.2020.230

Abstract ( 612 )   PDF (1291KB) ( 273 )   Save

References | Related Articles | Metrics

The subdivision Q-neighborhood vertex corona G1□·_QG2 of two regular graphs G₁ and G₂ for regular is the graph obtained from vertex disjoint union of Q(G1) and |V(G1)| copies of G₂, and by joining the neighbors of the i^th vertex of |V(G1)| to every vertex in the i^th copy of G₂; the subdivision Q-neighborhood edge corona G1□— _QG2 is the graph obtained from vertex disjoint union of Q(G1) and |I(G1)| copies of G₂, and by joining the neighbors of the i^th vertex of I(G1) to every vertex in the i^th copy of G₂, where Q(G1) is obtained from G₁ by inserting a new vertex into every edge of G₁ and then joining by edges those pair of new vertices which lie on adjacent edges of G₁, and I(G1) is denoted by the set of such new vertices inserted in each edge of G₁. Based on above, the generalized characteristic polynomial and Φ-spectrum of G1□·_QG2 and G1□— _QG2 are determined, respectively. As an application, their normalized Laplacian spectrum are obtained. Besides, infinitely many pairs of Φ-cospectral mates are also constructed.

Stability and Turing instability of a predator-prey reaction-diffusion system with schooling behavior

ZHOU Yan, ZHANG Cun-hua

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  73-81.  doi:10.6040/j.issn.1671-9352.0.2020.496

Abstract ( 917 )   PDF (12267KB) ( 223 )   Save

References | Related Articles | Metrics

A predator-prey reaction-diffusion system of fish with schooling behavior is considered. By analyzing the distribution of the roots of the characteristic equation of the linearized system at the unique constant positive equilibrium solution in the complex plane in detail, the stability and the Turing instability of the unique positive equilibrium solution of the system are discussed. In addition,the obtained theoretical conclusions are verified numerically by MATLAB software.

Extractive-abstractive text automatic summary based on BERT-SUMOPN model

TAN Jin-yuan, DIAO Yu-feng, YANG Liang, QI Rui-hua, LIN Hong-fei

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  82-90.  doi:10.6040/j.issn.1671-9352.1.2020.047

Abstract ( 1808 )   PDF (2565KB) ( 501 )   Save

References | Related Articles | Metrics

Extractive summaries have poor readability and accuracy, while abstractive summaries also have deficiencies in coherence and logic. In addition, the traditional models of the two summary methods are often insufficient and inaccurate for the vector representation of text. In response to the above problems, this paper proposes an extractive-abstractive summary method based on BERT-SUMOPN model. The model obtains the text vector through the BERT pre-trained language model, then extracts the key sentences in the text using the extractive summary model, and finally inputs the obtained key sentences into the pointer-generation network, and carries out the model through the EAC loss function for end-to-end training, combined with the coverage mechanism to reduce duplication and obtain summary results. The experimental results show that the BERT-SUMOPN model has achieved good results on the BIGPATENT patent dataset, and the ROUGE-1 and ROUGE-2 indicators have been improved by 3.3% and 2.5% respectively.

Multi-label feature selection based on manifold structure and flexible embedding

ZHANG Yao, MA Ying-cang, YAND Xiao-fei, ZHU Heng-dong, YANG Ting

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  91-102.  doi:10.6040/j.issn.1671-9352.0.2020.588

Abstract ( 729 )   PDF (12299KB) ( 270 )   Save

References | Related Articles | Metrics

The linear regression model and manifold structure are combined to form a weak linear multi-label feature selection framework. Firstly, the least square loss function is used to learn the regression coefficient matrix; secondly, the label manifold structure is used to learn the weight matrix of data features; thirdly, L_2,1-norm is used to constrain the regression coefficient matrix and feature weight matrix, which can guide the sparsity and facilitate feature selection. In addition, an iterative updating algorithm with convergence is designed and proved to solve the above problems. Finally, the proposed method is verified on several classical multi-label datasets, and the experimental results show the effectiveness of the proposed algorithm.

Study on the identification of Corii Colla Asini by nuclear magnetic resonance spectroscopy fingerprint

DENG Shu-hong, GUO Chuan-en, JIANG Hong, NIE Lei

JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE). 2021, 56(7):  103-110.  doi:10.6040/j.issn.1671-9352.0.2021.002

Abstract ( 990 )   PDF (5432KB) ( 345 )   Save

References | Related Articles | Metrics

To establish the nuclear magnetic resonance(NMR)fingerprinting method and to identify the Corii Colla Asini(Ejiao)with NMR fingerprinting. The ¹H-NMR based on the composition of the Corii Colla Asini water extract was used to establish its nuclear magnetic resonance fingerprint method, and the validation of method was also carried out. ¹H nuclear magnetic resonance spectra were measured for samples from different sources to construct samples data sets. Principal component analysis(PCA)and interval partial least squares(i-PLS)were employed to conduct pattern recognition of samples and a discriminant model of Corii Colla Asini was accordingly built. The precision, stability and reproducibility of the Corii Colla Asini NMR fingerprinting method were good. The analysis results of PCA showed that there were certain categorical differences between samples from different sources, but it was difficult to distinguish the authentic and fake Corii Colla Asini samples. The i-PLS method was used to establish the discriminant model of Corii Colla Asini, which had a strong classification ability and could correctly discriminate the authentic Corii Colla Asini and its fake samples. The combination of the ¹H NMR fingerprint and pattern recognition could be used to identify the authenticity of Corii Colla Asini, which would provide a significant reference for the quality control and evaluation of Corii Colla Asini.