Recursive Partitioning Analysis Classification and Graded Prognostic Assessment for Non-Small Cell Lung Cancer Patients with Brain Metastasis:A Retrospective Cohort Study

Objective：To assess prognostic factors and validate the effectiveness of recursive partitioning analysis （RPA） classes and graded prognostic assessment （GPA） in 290 non-small cell lung cancer （NSCLC） patients with brain metastasis （BM）.Methods：From Jan 2008 to Dec 2009,the clinical data of 290 NSCLC cases with BM treated with multiple modalities including brain irradiation,systemic chemotherapy and tyrosine kinase inhibitors （TKIs） in two institutes were analyzed.Survival was estimated by Kaplan-Meier method.The differences of survival rates in subgroups were assayed using log-rank test.Multivariate Cox＇s regression method was used to analyze the impact of prognostic factors on survival.Two prognostic indexes models （RPA and GPA） were validated respectively.Results：All patients were followed up for 1-44 months,the median survival time after brain irradiation and its corresponding 95% confidence interval （95% CI） was 14 （12.3-15.8） months.1-,2-and 3-year survival rates in the whole group were 56.0%,28.3%,and 12.0%,respectively.The survival curves of subgroups,stratified by both RPA and GPA,were significantly different （P0.001）.In the multivariate analysis as RPA and GPA entered Cox＇s regression model,Karnofsky performance status （KPS） ≥ 70,adenocarcinoma subtype,longer administration of TKIs remained their prognostic significance,RPA classes and GPA also appeared in the prognostic model.Conclusion：KPS ≥70,adenocarcinoma subtype,longer treatment of molecular targeted drug,and RPA classes and GPA are the independent prognostic factors affecting the survival rates of NSCLC patients with BM.

Diagnosis of building energy consumption in the 2012 CBECS data using heterogeneous effect of energy variables:A recursive partitioning approach

Numerous previous literature has attempted to apply machine learning techniques to analyze relationships between energy variables in energy consumption.However,most machine learning methods are primarily used for prediction through complicated learning processes at the expense of interpretability.Those methods have difficulties in evaluating the effect of energy variables on energy consumption and especially capturing their heterogeneous relationship.Therefore,to identify the energy consumption of the heterogeneous relationships in actual buildings,this study applies the MOdel-Based recursive partitioning(MOB)algorithm to the 2012 CBECS survey data,which would offer representative information about actual commercial building characteristics and energy consumption.With resultant tree-structured subgroups,the MOB tree reveals the heterogeneous effect of energy variables and mutual influences on building energy consumptions.The results of this study would provide insights for architects and engineers to develop energy conservative design and retrofit in U.S.office buildings.

Role of Recursive Partitioning Analysis and Graded Prognostic Assessment on Identifying Non-Small Cell Lung Cancer Patients with Brain Metastases Who May Benefit from Postradiation Systemic Therapy

Background：The role ofpostradiation systemic therapy in non-small cell lung cancer （NSCLC） patients with brain metastasis （BM） was controversial.Thus,we explored the role of Radiation Therapy Oncology Group recursive partitioning analysis （RTOG-RPA） and graded prognostic assessment （GPA） in identifying population who may benefit from postradiation systemic therapy.Methods：The clinical data of NSCLC patients with documented BM from August 2007 to April 2015 of two hospitals were studied retrospectively.Cox regression was used for multivariate analysis.Survival of patients with or without postradiation systemic therapy was compared in subgroups stratified according to RTOG-RPA or GPA.Results：Of 216 included patients,67.1% received stereotactic radiosurgery （SRS）,24.1% received whole-brain radiation therapy （WBRT）,and 8.8% received both.After radiotherapy,systemic therapy was administered in 58.3% of patients.Multivariate analysis found that postradiation systemic therapy （yes vs.no） （hazard ratio [HR] =0.36 l,95% confidence interval [CI] =0.202-0.648,P =0.001）,radiation technique （SRS vs.WBRT） （HR =0.462,95% CI =0.238-0.849,P =0.022）,extracranial metastasis （yes vs.no） （HR =3.970,95% CI =1.757-8.970,P =0.001）,and Kamofsky performance status （〈70 vs.≥70） （HR =5.338,95% CI =2.829-10.072,P 〈 0.001） were independent factors for survival.Further analysis found that subsequent tyrosine kinase inhibitor （TKI） therapy could significantly reduce the risk of mortality of patients in RTOG-RPA Class IⅡ （HR =0.411,95% CI =0.183-）.923,P =0.031） or with a GPA score of 1.5-2.5 （HR =0.420,95% CI =0.182-0.968,P =0.042）.However,none of the subgroups stratified according to RTOG-RPA or GPA benefited from the additional conventional chemotherapy.Conclusion：RTOG-RPA and GPA may be useful to identify beneficial populations in NSCLC patients with BM ifTKIs were chosen as postradiation systemic therapy.

Finding roots of arbitrary high order polynomials based on neural network recursive partitioning method

This paper proposes a novel recursive partitioning method based on constrained learning neural networks to find an arbitrary number (less than the order of the polynomial) of (real or complex) roots of arbitrary polynomials. Moreover, this paper also gives a BP network constrained learning algorithm (CLA) used in root-finders based on the constrained relations between the roots and the coefficients of polynomials. At the same time, an adaptive selection method for the parameter d P with the CLA is also given. The experimental results demonstrate that this method can more rapidly and effectively obtain the roots of arbitrary high order polynomials with higher precision than traditional root-finding approaches.

Forecasting Flowering and Maturity Times of Barley Using Six Machine Learning Algorithms

Interior Alaska has a short growing season of 110 d.The knowledge of timings of crop flowering and maturity will provide the information for the agricultural decision making.In this study,six machine learning algorithms,namely Linear Discriminant Analysis(LDA),Support Vector Machines(SVMs),k-nearest neighbor(kNN),Naïve Bayes(NB),Recursive Partitioning and Regression Trees(RPART),and Random Forest(RF),were selected to forecast the timings of barley flowering and maturity based on the Alaska Crop Datasets and climate data from 1991 to 2016 in Fairbanks,Alaska.Among 32 models fit to forecast flowering time,two from LDA,12 from SVMs,four from NB,three from RF outperformed models from other algorithms with the highest accuracy.Models from kNN performed worst to forecast flowering time.Among 32 models fit to forecast maturity time,two models from LDA outperformed the models from other algorithms.Models from kNN and RPART performed worst to forecast maturity time.Models from machine learning methods also provided a variable importance explanation.In this study,four out of six algorithms gave the same variable importance order.Sowing date was the most important variable to forecast flowering but less important variable to forecast maturity.The daily maximum temperature may be more important than daily minimum temperature to fit flowering models while daily minimum temperature may be more important than daily maximum temperature to fit maturity models.The results indicate that models from machine learning provide a promising technique in forecasting the timings of flowering and maturity of barley.

Identification of multi-target anti-cancer agents from TCM formula by in silico prediction and in vitro validation

Cancer is a complex disease associated with multiple gene mutations and malignant phenotypes,and multi-target drugs provide a promising therapy idea for the treatment of cancer.Natural products with abundant chemical structure types and rich pharmacological characteristics could be ideal sources for screening multi-target antineoplastic drugs.In this paper,50 tumor-related targets were collected by searching the Therapeutic Target Database and Thomson Reuters Integrity database,and a multi-target anti-cancer prediction system based on mt-QSAR models was constructed by using naïve Bayesian and recursive partitioning algorithm for the first time.Through the multi-target anti-cancer prediction system,some dominant fragments that act on multiple tumor-related targets were analyzed,which could be helpful in designing multi-target anti-cancer drugs.Anti-cancer traditional Chinese medicine(TCM)and its natural products were collected to form a TCM formula-based natural products library,and the potential targets of the natural products in the library were predicted by multi-target anti-cancer prediction system.As a result,alkaloids,flavonoids and terpenoids were predicted to act on multiple tumor-related targets.The predicted targets of some representative compounds were verified according to literature review and most of the selected natural compounds were found to exert certain anti-cancer activity in vitro biological experiments.In conclusion,the multi-target anti-cancer prediction system is very effective and reliable,and it could be further used for elucidating the functional mechanism of anti-cancer TCM formula and screening for multi-target anti-cancer drugs.The anti-cancer natural compounds found in this paper will lay important information for further study.

Nonlinear Program Construction and Verification Method Based on Partition Recursion and Morgan's Refinement Rules

The traditional program refinement strategy cannot be refined to an executable program,and there are issues such as low verification reliability and automation.To solve the above problems,this paper proposes a nonlinear program construction and verification method based on partition recursion and Morgan’s refinement rules.First,we use recursive definition technique to characterize the initial specification.The specification is then transformed into GCL(Guarded Command Language)programs using loop invariant derivation and Morgan’s refinement rules.Furthermore,VCG(Verification Condition Generator)is used in the GCL program to generate the verification condition automatically.The Isabelle theorem prover then validates the GCL program’s correctness.Finally,the GCL code generates a C++executable program automatically via the conversion system.The effectiveness of this method is demonstrated using binary tree preorder traversal program construction and verification as an example.This method addresses the problem that the construction process’s loop invariant is difficult to obtain and the refinement process is insufficiently detailed.At the same time,the method improves verification process automation and reduces the manual verification workload.

A Formal Method for Developing Algebraic and Numerical Algorithms

The development of algebraic and numerical algorithms is a kind of complicated creative work and it is difficult to guarantee the correctness of the algorithms. This paper introduces a systematic and unified formal development method of algebraic and numerical algorithms. The method implements the complete refinement process from abstract specifications to a concrete executable program. It uses the core idea of partition and recursion for formal derivation and combines the mathematical induction based on strict mathematical logic with Hoare axiom for correctness verification. This development method converts creative work into non-creative work as much as possible while ensuring the correctness of the algorithm, which can not only verify the correctness of the existing algebraic and numerical algorithms but also guide the development of efficient unknown algorithms for such problems. This paper takes the non-recursive implementation of the Extended Euclidean Algorithm and Horner's method as examples. Therefore, the effectiveness and feasibility of this method are further verified.