a statistical package based on pnuts
play

A Statistical Package Based on Pnuts Junji NAKANO The Institute of - PowerPoint PPT Presentation

A Statistical Package Based on Pnuts Junji NAKANO The Institute of Statistical Mathematics Takeshi FUJIWARA The Graduate University for Advanced Studies Yoshikazu YAMAMOTO Tokushima Bunri University Ikunori KOBAYASHI Tokushima Bunri University


  1. A Statistical Package Based on Pnuts Junji NAKANO The Institute of Statistical Mathematics Takeshi FUJIWARA The Graduate University for Advanced Studies Yoshikazu YAMAMOTO Tokushima Bunri University Ikunori KOBAYASHI Tokushima Bunri University JAPAN

  2. Overview � We are developing Jasp (Java based statistical processor) system � Outline of my talk 1. Motivations of making Jasp 2. Jasp language 3. Data based GUI 4. Client/server (C/S) approach and distributed computing 5. Foreign language interface 6. Conclusion 2

  3. 1. Motivation of making Jasp Existing statistical systems � We have a lot of statistical systems, which � have long history � are reliable � are sophisticated by adopting new technologies � GUI � Internet � are not so expensive (some are free!) � However, we want to have new one. Why? 3

  4. 1. Motivation of making Jasp Reasons for new system (1) � We need fully controllable system � Basic structure may be changed for realizing new ideas in statistical computing � We have to understand details of the system � We want to be free from the original author(s) 4

  5. 1. Motivation of making Jasp Reasons for new system (2) � Need for newly designed system � In many available systems, new functions were added on their original design � As computer has been changing rapidly, new design will be better for new technologies 5

  6. 1. Motivation of making Jasp Reasons for new system (3) � We (ISM) needs new statistical system � Many Fortran programs exist in ISM � They are not easy to use � Development of Data mining � Requirement for general purpose system � Statistical model using both mathematics and algorithm � We should have know-how for making large system 6

  7. 2. Jasp language Tool we selected: Java language (1) � Java adopts many new technologies in a well organized way � Platform independent � Purely object oriented � Good libraries � Network (TCP/IP) � GUI support and Graphics � Interface to other languages (Java Native Interface, JNI) 7

  8. 2. Jasp language Tool we selected: Java language (2) � Advanced network functions � Web based execution (Applet) � Using remote objects (Remote Method Invocation, RMI) � Security � Demerits � slow speed � politically unstable � Java is easy to use, at least compared with before � Merits are bigger than demerits 8

  9. 2 .Jasp language Statistical language � Function based languages (e.g. S, XploRe) � are intuitive and flexible � are not good at bundling similar notions � Object oriented languages (e.g. Java) � are good at arranging notions � are not easy to use tentatively � Both function based and object oriented abilities are required � Function based abilities are for interactive use � Object oriented abilities are for reusing programs 9

  10. 2. Jasp language Jasp is based on Pnuts � Pnuts is a function based script language interpreter written in Java � Simple syntax � Easy access to Java class � Built-in language extension facility, with examples for object oriented syntax � Pnuts -> Java translator � Source is available freely � http://javacenter.sun.co.jp/pnuts/ 10

  11. 2. Jasp language Statistical parts we used � “ Jampack ” for matrix manipulation � ftp://math.nist.gov/pub/Jampack/Jampack/Abo utJampack.html � “ Ptplot ” for graphics � http:// ptolemy.eecs.berkeley.edu/java/ ptplot/ � “ Colt ” for statistical distributions and random numbers � http://nicewww.cern.ch/hoschek/colt/ 11

  12. 2. Jasp language Example: Jasp functions − = 1 b ( X ' X ) X ' y function ols(y, x){ coeff = (x.trans * x).inv * x.trans * y return coeff } y = ˆ Xb function forecast(x, b){ y_hat = x * b return h_hat } 12

  13. 2. Jasp language Example: Jasp class jaspclass LinearRegression { method LinearRegression(y, x){ // constructor this.beta = ols(y,x) // slot this.forecast = forecast(x,this.beta) // slot } function ols(y, x){ coeff = (x.trans * x).inv * x.trans * y return coeff } function forecast(x, b){ y_hat = x * b return h_hat } } Note that functions are not modified 13

  14. 3. Data based GUI Graphical user interface � Recently, almost all software products are operated by GUIs � Most statistical systems also have GUIs � Show lists of functions as menus � Generate beautiful graphs by mouse operations 14

  15. 3. Data based GUI Character use interface � We need to manipulate data and models in complex manners � These operations are difficult to be defined by GUI � They are properly described by programs � CUIs are easy to use for professional or frequent users 15

  16. 3. Data based GUI Mixed user interface � Both GUI and CUI are required for statistical systems � We propose “ mixed user interface ” � GUI and CUI can be used almost independently � They can be used together seamlessly and alternatively 16

  17. 3. Data based GUI The history of the analysis � We need to do trials and errors to get appropriate models for data � Recording the history of the analysis is important � We implement it in our GUI as data based way 17

  18. 3. Data based GUI User interface of Jasp � We develop the user interface of Jasp for realizing an example of the mixed user interface � Jasp has two windows for GUI and CUI � Operations on one window are automatically recorded on the other window in appropriate forms 18

  19. 3. Data based GUI Roles of CUI � CUI is an environment for writing and executing a program in Jasp language � We can edit functions on the upper window � We can execute commands on the lower window and have results as characters 19

  20. 3. Data based GUI The CUI window Pull-down menus Buttons Editor area Input and output area 20

  21. 3. Data based GUI Roles of GUI � GUI window is another environment for operating Jasp mainly using mouse � All the functions and statistical objects are listed on the GUI window and can be used by mouse operations � Data are displayed as icons which are arranged to express the history of the analysis � Almost all operations can be performed through the GUI window � Exceptions: Defining new functions and new statistical objects 21

  22. 3. Data based GUI The GUI window Pull-down menu Statistics Icons Graphs Popup menu Functions, methods 22

  23. 3. Data based GUI Pop-up menu � Situation sensitive menu � Menu items are arranged by specified object and its state � Executed operations depend on specified object and its state � Menu items are applicable methods of selected object 23

  24. 3. Data based GUI Executing functions and methods � We can execute functions and methods in the lower left GUI window by mouse click � If the function or method requires some arguments, the window for specifying them are displayed � We can drag and drop icons on the upper left GUI window to the arguments of the function or method 24

  25. 4. C/S approach and distributed computing Client/server approach of Jasp � Jasp system consists of server and client � The user interface runs as the client program � Jasp client can be invoked both as an application and as a Java applet from Web browsers � Jasp runs on many platforms supported by Java virtual machines 25

  26. 4. C/S approach and distributed computing Client/server of Jasp GUI Calculation Message sending and by RMI program CUI Server Client 26

  27. 4. C/S approach and distributed computing Distributed computing in statistics � Need for distributed computing � We can use many powerful computers connected by network � Computer intensive statistical techniques are popular � simulation, resampling � maximization of complex likelihood function � Existing distributed computing technologies are not easy to use for statistician � MPI (Message Passing Interface) � PVM (Parallel Virtual Machine) 27

  28. 4. C/S approach and distributed computing Functions for distributed computing in Jasp � remotes(commands) � commands: ArrayList of commands � Execute commands simultaneously � remote(command,server) � Execute command on server � send(var1,var2,server) � var1: variable on local computer � var2: variable on server � Execute “ var2 = var1 ” on server 28

  29. 4. C/S approach and distributed computing Example of Distributed computing cmds = ArrayList() cmds.add(“mean1 = ranMean(50000)”) cmds.add(“mean2 = remote(¥”ranMean(50000)¥”,¥”Serv2¥”) cmds.add(“mean3 = remote(¥”ranMean(50000)¥”,¥”Serv3¥”) remotes(cmds) mean = (mean1 + mean2 + mean3) / 3 ranMean(n): Function to generate n random numbers and return their mean value 29

  30. 5. Foreign language interface Foreign language interface � We have many programs written in Fortran, C, and C+ + , for example, TIMSAC (TIMe Series Analysis and Control) package is a time series analysis program written in Fortran at ISM � Jasp has interface to such programs and use them in the same way as other Java libraries � We implement this using JNI(Java Native Interface) 30

  31. 5. Foreign language interface JNI(Java Native Interface) � JNI is Java API(Application Programming Interface) to interface Java with other language (Native code) � Using JNI, we can call native code from Java (and call Java code from native code) JNI Java code Native code � This is useful to reuse exiting programs and to increase processing speed 31

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend