Increasing recommended effectiveness with markov chains and purchase intervals

Recommendation system is an important component of many websites and has brought huge economic benefits and challenges for online shoppers and e-commerce companies. Existing recommendation systems focus on producing a list of products which users may be interested to purchase, while overlooking the purchase chain and temporal diversity which may increase the likelihood of a purchase decision. In this paper, we propose to utilize the Markov chain to track the chain of users’ purchase behaviors and utilize the purchase intervals to improve the temporal diversity for e-commerce recommender. We design and implement several algorithms and integrate these into our recommendation model. We evaluate our system on a real-world e-commerce dataset. Experimental results demonstrate that our approach significantly improves the accuracy, conversion rate and temporal diversity compared to the state-of-the-art algorithms.     

PARAMETER ESTIMATION IN EXPONENTIAL MODELS

Abstract. In this paper the problems of parameter estimation and order determination of an exponential (EX) model are studied in the time domain. In order to estimate the parameters, the parameter equations of an EX model are given in terms of the autocorrelation function, which is similar to the Yule-Walker equations of an autoregressive moving-average model. Estimates of parameters are obtained with the aid of the parameter equations and theorems are proved relating the convergence rate and asymptotic distribution of the estimates. We present two kinds of methods for estimating the order and prove that the estimates of the order are consistent.     

C-19 diacid-based polyamides

The preparation and properties of polyamide hot-melt adhesives derived from Koch C-19 diacid are described. They are comparable to those of commercial C-36 dimer acid-based polyamides for bonding leather and metals and somewhat superior to those of polyamides made from C-19 diacid derived from a hydroformylation process. Presented at the AOCS annual meeting, Toronto, May 1982.     

耐碱玻璃纤维砖骨料混凝土性能正交试验研究

以水胶比、粉煤灰掺量、砖骨料取代率、耐碱玻璃纤维体积掺量和耐碱玻璃纤维长度作为试验因素,通过设置五因素三水平正交试验,研究耐碱玻璃纤维二次破碎砖骨料混凝土28 d的立方体抗压强度以及28 d劈裂抗拉强度两种评价指标的变化规律.结果 表明:两种评价指标受水胶比、粉煤灰掺量、耐碱玻璃纤维体积掺量影响较大,二次破碎砖骨料取代率对劈裂抗拉强度影响较小,耐碱玻璃纤维长度对二者影响均不显著,综合考虑配置耐碱玻璃纤维砖骨料混凝土的最佳因素水平方案为水胶比0.4、砖骨料50％、粉煤灰10％、耐碱玻璃纤维体积掺量和长度分别为0.1％和16 mm.此结论可为耐碱玻璃纤维再生砖骨料新型混凝土工程使用提供试验依据.     

基于计算机测控技术的转矩流变仪的研制

介绍了基于计算机测控技术的聚合物转矩流变仪的研制，该仪器具有优良的性能指标，强大的数据文件处理能力、测试功能扩展能力，是高分子材料研究和开发的理想工具。     

生物发酵过程中SSVIP-NNI软测量方法

生物发酵过程中关键生化变量难以直接检测,提出了一种利用基于敏感性与特异性的变量投影重要性(SSVIP)方法优化神经网络逆系统(NNI)的软测量方法。根据逆系统理论建立软测量模型,采用VIP变量优选方法,对软测量模型中的辅助变量进行优化。为了进一步提高VIP方法优选变量的性能,利用模型敏感性与特异性的概念,重新定义了VIP筛选阈值,确定最优的发酵过程软测量模型的变量。构造神经网络近似最优逆系统软测量模型,实现对发酵过程中关键过程变量的估计。通过Pensim仿真平台进行实验研究,结果表明经过辅助变量优选的神经网络逆系统软测量模型具有更高的估计精度和泛化能力。     

固体超强酸S_2O_8~(2-)/Fe_2O_3—CoO催化合成乙酸异戊酯

制备固体超强酸催化剂S2O82-/Fe2O3—CoO,并用于乙酸异戊酯合成反应。用该催化剂通过正交试验合成乙酸异戊酯的最佳条件为:醇酸物质的量比为1.5,催化剂用量为0.8 g(以0.2 mol乙酸为准),带水剂用量为10 mL,反应时间为2.5 h,其酯化率可达98%以上。     

动态格点搜索的改进算法用于团簇结构快速优化

结合高效的动态格点搜索(DLS)算法与扰动操作(Perturbation Operation)提出一种新的改进方法(DLS-PO),用于确定团簇的最低能量结构。针对一个特定构型,DLS算法总能给出其对应搜索空间的最规则结构。然而,一次失败的DLS优化将消耗大量的运算资源。为此,采取原子移动和结构旋转的扰动操作成功地改变了构型,再结合后续的DLS操作,提高了优化效率。将该算法用于原子数高达309的Lennard-Jones团簇及100原子NP-B函数铝团簇的结构优化。优化结果显示相比于DLS算法,DLS-PO算法更为高效。     

边缘氢化石墨烯片层结构和光谱性质的密度泛函理论研究

以并3-6苯环为氢化石墨烯片层模型,采用量子化学密度泛函理论方法(DFT)在B3LYP/6-31G(d,p)水平上进行了几何结构全优化,讨论了分子结构、能量、前线分子轨道等性质的变化规律。在得到化合物基态稳定构型的基础上,运用含时密度泛函理论(TD-DFT)计算了电子吸收光谱的性质。计算结果显示边缘氢化石墨烯片层均为平面结构,线型扶手椅型结构能量最低,结构最稳定。随着石墨烯片层环数的增加HOMO轨道能量增加,LUMO轨道能量降低,能隙能量降低,最大吸收波长明显红移。     

橡胶注射过程中压力变化的研究

对橡胶注射过程中的压力变化进行了理论与实验研究。充模流动模型的压力、能量控制方程分别为：选用高斯方法求解总体有限元方程。实验与计算分析对比证明．本文选用的数学模型和计算方法所得结果与实测结果基本吻合、该模型可用以代替大量的实验，对橡胶充模过程进行模拟，并用Ｇｒａｐｈｅｒ软件对数值结果进行处理，以图表直观地表达出各种因素对充模过程的影响。