MLwiN 3.05 多層次模型軟體-統計分析軟體/新永資訊有限公司

MLwiN 3.05 多層次模型軟體

MLwiN 3.05 多層次模型軟體

  • MLwiN 3.05 多層次模型軟體
  • 編號
  • 類別
  • 介紹
  • 價格

MLwiN 3.05 多層次模型軟體

User interface
The user interface has been completely updated, making it more consistent and easy to use. This also means that a 64-bit version of the GUI is available, allowing larger dataset plus possible speed improvements.

Ease of use features
It is now possible to display multiple graph windows and MCMC diagnostics windows at the same time, making comparisons easier.

When pasting data with a preview it is now possible to review the whole data being pasted prior to committing.

The facility to define commonly used test structures has been added to the “intervals and tests” window. This basic setup facility allows straightforward testing of the fixed and random parts of the model; this is done without affecting the full flexibility of the procedure. The Intervals and Tests window has also been modified to make it clearer what needs to be input by the user and what is provided the program when results are calculated.

It is now possible to directly call OpenBUGS or WinBUGS, automatically bringing back the parameters chains on completion.

Information regarding the number of higher levels in use for the current model is now displayed on the equations window to alert users to the structure being analysed.

New Windows
A window for setting up 3D graphs has been added
A window for viewing and editing the current content of boxes has been added
A window for viewing and editing the current content of strings has been added

Double-precision columns
Column values in the worksheet are now stored in double precision. This will improve compatibility when saving or loading from
other packages, however there may be slight differences in model results compared to those from previous versions. This is particularly likely to happen where intermediate results are stored on
the worksheet, for example in discrete models or the starting
residuals when running MCMC models.

New default random number generator
The default random number generator has changed from the
3-seed Wichmann-Hill to Mersenne Twister as the former is
considered to have an inadequate cycle length . The old
random number generator is still the default for worksheets
saved in previous MLwiN versions to ensure consistency of
results. The random number generator can be set using the
RNGV command, which has the following values:

MLwiN worksheets with a format corresponding to version three or higher save the current state of the Mersenne Twister random number generator in the worksheet, so results will be consistent between saving and loading the worksheet. If the worksheet is
saved to an older version then this information will be lost and
the generator state will not persist.

Changes to model comparison functionality
When storing model results for MCMC estimation summary
statistics for the chains are now calculated for correlations as
well asvariances and covariances. The values to display can
then bespecified when operating the “Manage stored models
” window. Note,calculating this information requires the parameter chains and so is only available for models stored with MLwiN 3 or later. Thesechanges allow the storing and display of credible intervals for correlations between random effects as well as the covariances when MCMC estimation has been used. As these may
take quite a long time to calculate from the full length of the
chains, the user is kept informed of what is being calculated.

New DTOC command
It is sometimes useful (for example when plotting data by groups) to convert unique values of a continuous variable into discrete categories. The DTOC command will create a new column where each unique values is assigned a category number from zero,
with the original values assigned to the category labels.


MLwiN is developed and tested on Windows 10. It may work on previous versions of Windows, however we are unable to test and therefore support this.
32 Mb RAM
100Mb hard disk.

MLwiN 3.05 多層次模型軟體

用戶界面已完全更新,使其更加一致且易於使用。這也意味著 64 位版本的 GUI 可用,允許更大的數據集以及可能的速度改進。

現在可以同時顯示多個圖形窗口和 MCMC 診斷窗口,使比較更容易。



現在可以直接調用 OpenBUGS 或 WinBUGS,完成後自動帶回參數鏈。


添加了用於設置 3D 圖形的窗口

工作表中的列值現在以雙精度存儲。這將提高從其他包保存或加載時的兼容性,但是與以前版本的模型結果相比,模型結果可能略有不同。在中間結果存儲在工作表中的情況下尤其可能發生這種情況,例如在離散模型中或運行 MCMC 模型時的起始殘差。

默認隨機數生成器已從 3-seed Wichmann-Hill 更改為 Mersenne Twister,因為前者被認為循環長度不足。舊的隨機數生成器仍然是以前 MLwiN 版本中保存的工作表的默認設置,以確保結果的一致性。可以使用 RNGV 命令設置隨機數生成器,該命令具有以下值:

將該值設置為十(新工作表的默認值)或更高將使用內置的 C++ 概率分佈函數,而不是內置於 MLwiN 中的函數。

具有與版本 3 或更高版本對應的格式的 MLwiN 工作表將 Mersenne Twister 隨機數生成器的當前狀態保存在工作表中,因此保存和加載工作表之間的結果將保持一致。如果工作表保存為舊版本,則此信息將丟失並且生成器狀態將不會持續存在。

當存儲 MCMC 估計的模型結果時,鏈的匯總統計現在計算相關性以及方差和協方差。然後可以在操作“管理存儲模型”窗口時指定要顯示的值。請注意,計算此信息需要參數鏈,因此僅適用於使用 MLwiN 3 或更高版本存儲的模型。當使用 MCMC 估計時,這些更改允許存儲和顯示隨機效應之間的相關性以及協方差的可信區間。由於從鏈的全長計算這些可能需要相當長的時間,因此用戶會被告知正在計算的內容。

新的 DTOC 命令
將連續變量的唯一值轉換為離散類別有時很有用(例如在按組繪製數據時)。DTOC 命令將創建一個新列,其中每個唯一值都被分配一個從零開始的類別編號,原始值分配給類別標籤。


Analyse-it 5.56 方法評估分析軟體

Analyse-it 是一個方法評估分析軟體,檢驗科或者那些需要檢驗、核實分析診斷數據方法的研究員使用


Analyse-it 5.56 方法評估分析軟體

XpertRule® Miner 資料採礦軟體

XpertRule Miner所採用之資料採礦(data mining)是最近學術界與產業界十分重視的技術,它是藉由資訊技術從大量資料裡粹取出資訊與結果之間的關聯性。企業可以用它得到顧客與購買產品之間的相關資訊。現今大多數企業在分析他們的客戶資料時,採取傳統統計方法所做出的結果,仍具有改進的空間。


XpertRule® Miner 資料採礦軟體

Scientific Computing Associates (SCA) 8.1 時間數列統計軟體

SCA System 8.1 版分為多個版本,旨在滿足在時間序列分析和預測應用程序方面具有不同能力要求和專業知識的用戶的需求。SCAB34S 小程序和 B34S ProSeries 計量經濟學系統中提供了強大的 GARCH、數據挖掘、預測建模和高級計量經濟學建模附加組件。


Scientific Computing Associates (SCA) 8.1 時間數列統計軟體