Meta-analysis is a method of collecting, merging and quantitative statistical analysis of research results from different sources according to a clear search strategy, selection and screening of literature standards, and strict evaluation methods for a certain scientific research problem. It first appeared in "evidence-based medicine". ", has been widely used in agriculture and forestry ecology, resource environment and other aspects. The R language has a complete and effective data processing, statistical analysis and storage mechanism, which can directly analyze and display the data. The command format is simple and the results are highly readable. It includes many software packages for Meta analysis, which is the best choice for Meta integration analysis and evaluation. effective platform. Research hotspot changes from bibliometric analysis, search for scientific issues, R-Meta multi-method full-process analysis and Meta advanced drawing, multi-level hierarchical nested model construction and Meta regression diagnosis, Bayesian network, MCMC parameter optimization and uncertainty Analysis, six methods of Meta data missing value processing and result reliability analysis, Meta weighted machine learning and non-linear Meta analysis, etc., each topic, each part combined with multiple typical case practices, was well received by many students.
| main content |
|
| Topic 1 Topic Selection and Retrieval for Meta-analysis |
1. Topic selection and literature search for meta - analysis 1) What is Meta-analysis 2) Topic selection strategy for meta-analysis 3) Accurate retrieval strategy, how to retrieve complete and accurate retrieval 4) Literature management and cleaning, how to formulate literature inclusion and exclusion criteria 5) Literature data acquisition skills, research topic exploration and scientific problem proposal 6) Bibliometric analysis of CiteSpace, VOSViewer, R bibliometrix and analysis of research hotspots
|
| Topic 2 Meta analysis and R language data cleaning and statistical methods |
2. Commonly used software for Meta analysis/R language basis and statistical basis 1) The advantages of R language for Meta analysis and its classic case application in " Nature " and " Science " 2) R language basic operation and data cleaning method 3) Statistical basis and calculation of commonly used statistics (sd\se\CI) , three major tests (T test, chi-square test and F test) 4) Similarities and differences between traditional statistics and Meta analysis 5) Explanation of commonly used packages and related plug-ins for Meta analysis in R language From self-programming calculation to calling Meta packages ( meta, metafor, dmetar, esc, metasens, metamisc, meta4diag, gemtc , robvis , netmeta , brms , etc.), the whole process analyzes how to perform meta calculation, meta diagnosis, Bayesian meta, mesh meta, subgroup analysis, meta regression and graphing .
|
| Topic Three R language Meta analysis and graphing |
3. Calculation of Meta effect value in R language 1) The process of Meta analysis in R language 2) Comparison of various meta effect value calculations, self-programming and calling functions lnRR, MD and SMD of continuous data RR and OR for categorical data 3) The use of R language meta package and metafor package 4) How to draw a beautiful forest map with R base package and ggplot2
|
| Topic 4 R language Meta regression analysis |
4. R language Meta analysis and mixed effect model (hierarchical model) construction 1) Meta分析的权重计算 2) Meta分析中的固定效应、随机效应 3) 如何对Meta模型进行统计检验和构建嵌套模型、分层模型（混合效应） 4) Meta回归和普通回归、混合效应模型的对比及结果分析 5) 使用Rbase和ggplot2绘制Meta回归图
|
| 专题五 R语言Meta诊断分析 |
5、R语言Meta诊断进阶 1) Meta诊断分析（t2、I2、H2、R2、Q、QE、QM等统计量） 2) 异质性检验及发表偏移、漏斗图、雷达图、发表偏倚统计检验 3) 敏感性分析、增一法、留一法、增一法、Gosh图 4) 风险分析、失安全系数计算 5) Meta模型比较和模型的可靠性评价 6) Bootstrap重采样方法评估模型的不确定性 7) 如何使用多种方法对文献中的SD、样本量等缺失值的处理
|
| 专题六 R语言Meta分析的不确定性 |
6、R语言Meta分析的不确定性 1) 网状Meta分析 2) 贝叶斯理论和蒙特拉罗马尔可夫链MCMC 3) 如何使用MCMC优化普通回归模型和Meta模型参数 4) R语言贝叶斯工具Stan、JAGS和brms 5) Bayesian Meta analysis and uncertainty analysis
|
| Topic Seven Application of Machine Learning in Meta-analysis |
7. Application of Machine Learning in Meta-analysis 6) The basis of machine learning and the advantages of Meta machine learning 7) Use of Meta Weighted Random Forest (MetaForest) 8) Use Meta machine learning and traditional machine learning to train and test big data in the literature 9) How to judge whether Meta machine learning uses random effects or fixed effects and the optimization of hyperparameters 10) Use Meta machine learning for driving factor analysis and partial independent analysis of PDP
|
| Topic Eight Discussion and Q&A |
1 exercise 2 Discussion and Q&A |
Link to the original text: The application of the core technology of meta-analysis in the whole process of R language