Mplus是一個統計建模軟體，為研究人員提供了一個靈活的工具來分析資料，提供了多種選擇，具有易於使用的圖形介面和展示資料分析結果的模型，估計和演算法的選擇。
為研究人員提供了一個靈活的工具來分析資料，提供了多種選擇，具有易於使用的圖形介面和展示資料分析結果的模型，估計和演算法的選擇。Mplus允許分析橫斷面和縱向資料，單層和多層資料，來自不同的母體的資料，無論可見或不可見的異質，以及包含缺失值的資料，censored, binary, ordered categorical (ordinal), unordered categorical (nominal), counts或這些變數類型的組合進行分析。Mplus對遺漏值、複雜的調查資料和多層次的資料也有特別的功能。此外，Mplus對在蒙地卡羅模擬研究的任何可產生和分析資料的模型具有強大功能。
Mplus模組框架繪製了潛在變量的統一主題。 Mplus模組框架的一般性來自連續和分類潛變量的獨特使用。連續的潛變量對應於未觀察到的構建體的因子，對應於個體發育差異的隨機效應，對應於分層數據中群體間係數變化的隨機效應，對應於未觀察到的存活時間異質性的弱化性，疾病和對應於缺失數據的潛在應答變量值。分類潛變量用於表示對應於同類群個體的潛類，對應於未觀察群體中的發育類型的潛在軌跡類，對應於未觀察群體的有限混合物的混合物成分以及對應於缺失數據的潛在響應變量類別。
The Mplus Modeling Framework
模組數據的目的是以簡單的方式描述數據的結構，以便可理解和解釋。實質上，數據的模組等同於指定變量之間的一組關係。下圖顯示了可以在Mplus中建模的關係型態。 矩形代表觀察到的變量。 觀察到的變量可以是結果變量或背景變量。 背景變量被稱為x; 連續和刪掉的結果變量被稱為y; 和二元的，有序的分類（有序的），無序的分類（名義）和計數結果變量被稱為u。 圓圈代表潛在變量。 連續的和分類的潛變量都是允許的。連續的潛變量被稱為f。 分類潛變量被稱為c。
圖中的箭頭表示變量之間的回歸關係。圖中允許但未具體顯示的回歸關係包括觀察結果變量，連續潛在變量和分類潛變量之間的回歸。對於連續的結果變量，使用線性回歸模型。對於審查的結果變量，使用審查（回溯）回歸模型，在審查點有沒有通貨膨脹。 對於二元和有序的分類結果，使用probit或logistic回歸模型。 對於無序分類結果，使用多項邏輯回歸模型。對於計數結果，使用泊松和負二項回歸模型，在零點處有沒有通貨膨脹。
Mplus中的模型可以包括連續的潛變量，分類潛變量或連續和分類潛變量的組合。在上圖中，橢圓A描述了只有連續潛變量的模型。 橢圓B描述僅具有分類潛變量的模型。完整的模組框架描述了連續和分類潛變量組合的模型。上圖中的部分和之間的部分，可以使用Mplus來評估描述個體水平（範圍內）和群集水平（之間）變化的多級模型。
時間序列分析
時間序列分析用於分析密集縱向數據，如採用生態瞬時評估，經驗抽樣方法，日常日記法和動態評估方法獲得的數據。這樣的數據通常具有大量的時間點，例如二至二百個。測量通常在時間上緊密相隔。在Mplus中，可以估計N = 1型和各種兩級和交叉分類的時間序列模型。這些包括單變量自回歸，回歸，交叉滯後，確認因子分析，項目響應理論和連續，二進制，有序分類（序數）或這些變量類型的組合的結構方程模型。使用貝葉斯分析（Asparouhov，Hamaker＆Muthén，2017）進行估計。
Mplus Base Program
The Mplus Base Program estimates regression, path analysis, exploratory and confirmatory factor analysis (EFA and CFA), structural equation (SEM), growth, and discrete- and continuous-time survival analysis models. In regression and path analysis models, observed dependent variables can be continuous, censored, binary, ordered categorical (ordinal), counts, or a combination of these variable types. In addition, for regression analysis and path analysis for non-mediating variables, observed dependent variables can be unordered categorical (nominal). In EFA, factor indicators can be continuous, binary, ordered categorical (ordinal), or a combination of these variable types. In CFA, SEM, and growth models, observed dependent variables can be continuous, censored, binary, ordered categorical (ordinal), unordered categorical (nominal), counts, or a combination of these variable types. Other special features include single or multiple group analysis; missing data estimation; complex survey data analysis including stratification, clustering, and unequal probabilities of selection (sampling weights); latent variable interactions and non-linear factor analysis using maximum likelihood; random slopes; individually-varying times of observation; non-linear parameter constraints; indirect effects; maximum likelihood estimation for all outcomes types; bootstrap standard errors and confidence intervals; Bayesian analysis and multiple imputation; Monte Carlo simulation facilities; and a post-processing graphics module.
Mplus Base Program and Mixture Add-On
The Mplus Base Program and Mixture Add-On contains all of the features of the Mplus Base Program. In addition, it estimates regression mixture models; path analysis mixture models; latent class analysis; latent class analysis with multiple categorical latent variables; loglinear models; finite mixture models; Complier Average Causal Effect (CACE) models; latent class growth analysis; latent transition analysis; hidden Markov models; and discrete- and continuous-time survival mixture analysis. Observed dependent variables can be continuous, censored, binary, ordered categorical (ordinal), unordered categorical (nominal), counts, or a combination of these variable types. Other special features include single or multiple group analysis; missing data estimation; complex survey data analysis including stratification, clustering, and unequal probabilities of selection (sampling weights); latent variable interactions and non-linear factor analysis using maximum likelihood; random slopes; individually-varying times of observation; non-linear parameter constraints; indirect effects; maximum likelihood estimation for all outcomes types; bootstrap standard errors and confidence intervals; automatic starting values with random starts; Bayesian analysis and multiple imputation; Monte Carlo simulation facilities; and a post-processing graphics module.
Mplus Base Program and Multilevel Add-On
The Mplus Base Program and Multilevel Add-On contains all of the features of the Mplus Base Program. In addition, it estimates models for clustered data using multilevel models. These models include multilevel regression analysis, multilevel path analysis, multilevel factor analysis, multilevel structural equation modeling, multilevel growth modeling, and multilevel discrete- and continuous-time survival models. In multilevel analysis, observed dependent variables can be continuous, censored, binary, ordered categorical (ordinal), unordered categorical (nominal), counts, or a combination of these variable types. Other special features include single or multiple group analysis; missing data estimation; complex survey data analysis including stratification, clustering, and unequal probabilities of selection (sampling weights); latent variable interactions and non-linear factor analysis using maximum likelihood; random slopes; individually-varying times of observation; non-linear parameter constraints; maximum likelihood estimation for all outcomes types; Bayesian analysis and multiple imputation; Monte Carlo simulation facilities; and a post-processing graphics module.
Mplus Base Program and Combination Add-On
The Mplus Base Program and Combination Add-On contains all of the features of the Mplus Base Program and the Mixture and Multilevel Add-Ons. In addition, it includes models that handle both clustered data and latent classes in the same model, for example, two-level regression mixture analysis, two-level mixture confirmatory factor analysis (CFA) and structural equation modeling (SEM), and two-level latent class analysis, multilevel growth mixture modeling, and two-level discrete- and continuous-time survival mixture analysis. Other special features include missing data estimation; complex survey data analysis including stratification, clustering, and unequal probabilities of selection (sampling weights); latent variable interactions and non-linear factor analysis using maximum likelihood; random slopes; individually-varying times of observation; non-linear parameter constraints; maximum likelihood estimation for all outcomes types; Bayesian analysis and multiple imputation; Monte Carlo simulation facilities; and a post-processing graphics module.
Note: Version 7.31 is not available for Windows XP/ME/Vista, Mac OS X 10.6 Snow Leopard, Mac OS X 10.7 Lion and 32-bit Linux operating systems.
Mplus Version 7.31 is available for 32-bit and 64-bit Windows, 64-bit Mac OS X and 64-bit Linux operating systems.
Memory requirements for Mplus (both RAM and virtual memory) depend on the type of analysis and the amount of data. For example, analyses using numerical integration with large samples can be computationally heavy and may require a large amount of memory.
Mplus 32-bit on 32-bit Windows: The Windows 32-bit operating system limits all applications to a maximum of 2GB of total memory (RAM and virtual memory).
Starting with Mplus Version 4.21, Mplus is capable of using an additional 1 GB of total memory, for a maximum of 3 GB, on some Windows 32-bit operating systems. For Mplus to use the additional memory, the /3GB switch must be specified in Boot.ini. This switch allows applications like Mplus, that are “large address aware”, to use 1 GB of additional total memory above 2 GB.
Note that changes to the Boot.ini may have undesirable effects. Caution should be taken when making any changes. Muthen & Muthen is not responsible for any problems that may arise if such changes are attempted and does not provide support for any problems related to these changes.
Mplus 32-bit on 64-bit Windows: The Windows 64-bit operating system limits all 32-bit applications to a maximum of 2GB (or 4GB for applications like Mplus that are “large address aware”) of total memory (RAM and virtual memory). Note that the /3GB switch is not applicable to Windows 64-bit operating systems.
Mplus on 64-bit Windows: The Windows 64-bit operating system limits all 64-bit applications to a maximum of 8 terabytes (8000GB) of total memory (RAM and virtual memory).
Mplus on Mac OS X: Starting with Version 6.11, Mplus is available on Mac OS X. Mplus Version 7.1 for Mac OS X is available with an Mplus Editor. For previous versions of Mplus, some users have reported success installing Mplus on Intel-powered Mac using either Parallels Desktop or VMware Fusion software. Mplus support will not be provided for the installation and use of these software or the installation of Mplus using these software.
Mplus on Linux: Starting with Version 6.11, Mplus is available on the Linux operating systems from the command line. The Linux distributions that have been tested are Ubuntu, RedHat, Fedora, Debian, and Gentoo. Mplus for Linux does not run on the Sun Solaris operating system. The multiprocessing feature in Mplus for Linux is the standard multiprocessing feature available in Mplus for Windows. The Linux version does not currently have special cluster computing capabilities.
For previous versions of Mplus, some users have reported success using Mplus on Linux operating systems with the use of Wine. There was a problem running Version 5.2 with Wine, but the problem seems to be resolved starting with Version 6.1.
The PATH Environment Variable
The Run Mplus option in the Mplus Editor requires that the Mplus installation directory be listed in the PATH environment variable. The PATH environment variable must also be set at the system level for Mplus to work properly in a multiple user environment. The Mplus installation process will try to set the PATH environment variable at the system level.
If this attempt fails, problems may arise when using the Run Mplus option in the Mplus Editor or when trying to update Mplus. Users will be alerted that Mplus is unable to start when running Mplus or Mplus is not found when updating Mplus.
To fix this problem, the PATH environment variable can be set manually.
Instructions for Windows 7:
Mplus Temporary Files
Mplus uses intermediate files for controlling outputs. These files are created in the working directory. The working directory is usually the directory with the input file. There is one instance when it is not and that is when Mplus is first opened and the recent file list is used to open a previous input file. To ensure that the working directory is always the directory with the input file, always use the File -> Open/Save options for opening/saving an input file.
There is one other temporary file that is created by Mplus but not stored in the working directory. This file does not contain any specific results from the current run. The only information written to the file is the name of the output file. Mplus stores this file in the directory specified by the TEMP environment variable — a setting on most computers with the Windows operating system. There is no specific setting in Mplus to change the location of this directory. To change the location where Mplus stores this file, the TEMP environment variable must be changed. This will affect other programs that use the TEMP environment variable.
Instructions for Mplus Silent Installation for Windows
In order to perform a silent installation, the installer must be run once to get the response file used. The installer file must be started via the command line using the following instructions.
請留下完整資訊，以便我們提供精確的服務內容給你。