Abstracts


SIS ABSTRACTS


Responsible staff:

Svein Nordbotten            Joan C. Nordbotten
SVEIN NORDBOTTEN            Department of Information and Media Sciences
Birkelundsbakken 11A        University of Bergen
N-5231 Paradis              N-5020 Bergen,
NORWAY                      NORWAY

For some of the papers below and all papers published after year 2000, full-text versions are available.


Nordbotten, J.C. and Nordbotten,S. (1981): System Analysis and Design Applied to Statistical Systems. Invited Rice Lecture. Proceedings from the 43rd Session of the International Statistical Institue. Buenos Aires. 24p.

Abstract: A statistical system is a large intetgrated information system which commonly contains several automated subsystems. System analysis and design is an approach towards formalizing the analysis and design of information systems with the objective of improving the system performance by automation. In this paper we review the extent to which the methodologies of system analysis and design have been applied to the design of statistical systems and note the degree of understanding provided by their use. We maintain that continued and extended use of current system analysis and design methodologies will aid in the design of successful statistical systems.


Nordbotten, S. (1993): Statistical Meta-Knowledge and -Data. Statistical Journal of UN/ECE, Vol. 10, No.2, pp.101-111.

Abstract: Statistical meta-knowledge and -data requirements are described in this paper and the desired content of metadata in statistical systems are outlined. Finally, the needs for further efforts in developing components missing and for consolidating research results are emphazised.


Nordbotten, S. (1994) : International Statistical Organization - Comments to a Lecture by Dr. Zoltan E. Kennessey . In Kennessey, Z.E.: The Future of STATISTICS - An International Perspective. EUROSTAT/International Statistical Institute/Central Bureau of Statistics of The Netherlands. Voorburg. pp. 153-159.

Abstract: Why should we try to foresee the future, who will have needs for international statistics, in which form will they be demanded, what will the task of an international statistical system be, and finally, how should international statistics be organized?


Nordbotten, S. (1995): Editing Statistical Records by Neural Networks. Journal of Official Statistics, Vol. 11, No.4, pp. 391-411.

Abstract: Editing records to ensure quality in statistical surveys is an expensive and time consuming process. Since the introduction of computers in statistical processing, development and use of automated editing have been important objectives. In this article, editing is reformulated as a neural network problem. The use of such a network is demonstrated and a preliminary evaluation is presented.


Nordbotten, S. (1996): A Model for Automatic Coding from Boxes of Handwritten Characters. Proceedings of the 1996 Annual Research Conference, Bureau of the Census, US Department of Commerce, Washington D.C., pp. 791-809.

Abstract: In mail surveys, it is frequently necessary to ask the respondents to describe themselves by their placxe of birth, profession, education, etc. in written form. The answers must subsequently be coded by the statistical agency. Modern optical scanning equipment can transcribe the answers to machine-readable form. Because of distorted characters and unusual spelling, automatic coding may be a possible solution. This paper outlines a model for automatic handling of distorted characters and spelling errors. The model was tested on US city names by means of a prototype system. The system and the results are discussed in the paper.


Nordbotten, S. (1996): Neural Network Imputation Applied to the Norwegian 1990 Census Data. Journal of Official Statistics, Vol. 12, No.4, pp. 385-401.

Abstract: When adequate administrative registers are avaiable, population censuses can be carried out by supplementing the administrative data with data compiled by sample surveys. A problem is, however, to obtain reliable estimates from the survey for small areas and groups. This paper article results from empirical experiments with neural network models for imputing individual values for survey attributes utilizing the available administrative data. The imputed values were used to prepare estimates of proportions for a population and for smaller subgroups of the population. The data used in the experiment were from the 1990 Population Census in Norway.


Nordbotten, S. (1996): Small Area Statistics Based on imputations from Survey Data and Administrative Registers. SIS Memorandum.

Abstract: In many statistical surveys, estimates for small areas or populations cannot be released because of their low degree of accuracy. This paper reports on use of aggregated totals based on imputed/estimated values for individuals in small areas instead of totals estimated from a sample of individuals.


Nordbotten, S. (1996): Editing and Imputation by Means of Neural Networks. Statistical Journal of UN/ECE, Vol.13, No.2 , pp. 119-129.

Abstract: Editing and imputation of statistical data are possible because we take advantage of some prior knowledge about the type of statistical objects we investigate. The processes of editing and imputation are considered expensive parts of survey and census costs. This presentationdiscusses the nuse of neural network methodology to improve the efficiency of these processes. Two applications are discussed as demonstrations of the approach.


Nordbotten, S. (1996): Editing and Imputation by Means of Neural Networks. Chapter 10 in Alvey, W. and Jamerson, B. (eds.): Data Editing Workshop and Exposition. US Office of Management and Budget. Washington, DC. 1996. pp.213-242.

Abstract:Editing and imputation of statistical data are possible because we take advantage of some prior knowledge about the type of statistical objects we investigate. The processes of editing and imputation are considered expensive parts of survey and census costs. This presentationdiscusses the nuse of neural network methodology to improve the efficiency of these processes. Two applications are discussed as demonstrations of the approach.


Nordbotten, S. (1997): A Hybrid System for Automatic Coding of Handprinted Responses in Statistical Surveys . Neural Network World, Vol.7, No.2, pp191-203.

Abstract:In mail surveys, it is frequently necessary to ask the respondents to describe themselves by their placxe of birth, profession, education, etc. in written form. The answers must subsequently be coded by the statistical agency. Modern optical scanning equipment can transcribe the answers to machine-readable form. Because of distorted characters and unusual spelling, automatic coding may be a possible solution. This paper outlines a model for automatic handling of distorted characters and spelling errors. The model was tested on US city names by means of a prototype system. The system and the results are discussed in the paper.


Nordbotten, S.(1996): Presentation at the UN/ECE Work Session on Statistical Data Editing in Voorburg. The Netherlands, November 1996. UN/ECE Secretariat, Geneva.

Abstract: This paper outlines four cases of neural network application for improving the quality of statistics by editing and imputating individual data records. Some empirical experiences from the investigations are presented with tentative conclusions.


Nordbotten, J.C. and Nordbotten. S.(1997): Information Dissemination Using Hypertext Exhibits. Norsk Tidsskrift for Biblioteksforskning Nr. 10. 1997. pp. 76-90.

Abstract:We 've studied the usage characteristics of a small, multi theme,hypertext exhibit of social science research projects ranging from current financial networks in Norway tothe statistical system of the Incas. The exhibit was constructed as part of the University of Bergen's 50th anniversary celebration, which opened in August 1996 at the Museum for Natural Sciences. It was installed using a WebSite server and Netscape browser on a stand-alone PC with a touch screen. An English, www version of the exhibit is scheduled for release in June 1997.


Nordbotten, S. (1997): Models of Compex Human Screening and Correcting of Social Data. Computers in Human Behavior, Vol.13, No.4, pp.487-503.

Abstract: National statistical agencies spend extensive resources in continuous collection of huge amount of data on individual persons, households, business activities, etc., to serve the information needs of national and international government and private users. Substantial parts of their budgets are consumed in checking and improving the quality of the data collected. Because of their complexities, these tasks have traditionbally been carried out by specialists. To save both processing time and resources and to improve the servicing data requests, these tasks have been given high priority. The present paper outlines research carried out in Norway on using the neural network paradigm for more efficient data checking and improvement in large-scale data masses.


Nordbotten, S.(1996): Metrics for the Quality of Editing, Imputation and Prediction. UN/ECE Work Session on Statistical Data Editing, Prague, The Czech Republic, October 1996. UN/ECE Secretariat, Geneva.

Abstract: The quality of statistics will always be uncwertain. To serve the users, quality declarations are desirable. The aim of the editing process is to improve the quality in statistical products. In this paper, metrics for predicting the qualities of statistical products are proposed. The quality measures can be compared with quality requirements and be the basis for quality declarations. The reliability of the proposed metrics are illustrated by empirical tests.


Nordbotten, S. and Nordbotten, J.C.(2000): Perception of Statistical Presentations - Investigated by means of Internet Experiments. Proceedings of the 34th Hawaii International Conference on System Sciences January 3-6 2001. Institute of Electrical and Electronic Engineers.

Abstract: One of the main responsibilities of a national statistical office is to disseminate statistical information about the state and development of its country. The information may be a responses to requests for statistical publications or for general distribution in newspapers, radio, television, etc. The information disseminated may be in the form of statistical tables, graphical charts, verbal or vocal presentations depending on the public to which the presentation is aimed.

It has been assumed that statistical tables are particularly useful and needed for detailed analysis and research while graphical charts are more suitable for fast visual orientation of mainlines within the area on which information is requested. The subject of this proposal is the graphical chart as a presentation form and the problem to be investigated is how well different chart formats are perceived and correctly interpreted. Our interest is in the general public receiving statistical news by newspapers, television and other visualizing media. Typical for these 'statistical users' is that they have not requested any particular information and may not have any particular interest in the statistical fact presented. They observe and store the information in order to implicitly build opinions about their menvironment.

The justification for the project is that considerable resources are spent to disseminate statistics from the statistical bureaus in order to prepare the citizens to make more informed decisions.


Nordbotten, S. (1998): New methods of Editing and Imputation. Agriculture Statistics 2000. An International Conference on Agriculture Statistics. Washington, DC. March 18-20, 1998.

Abstract: Editing of data collected for preparation of statistics is a time and resourcer consuming process. This paper presents experiments with artificial neural networks as a potential tool for increasing the effectiveness of statistical editing and imputation. To maintain accuracy in resulting statistics, the possibility of deriving reliable accuracy predictions is also discussed.

A full draft is available at Agriculture Statistics 200.


Nordbotten, S. (1998):Estimating Population Proportions from Imputed Data. Computational Statistics & Data Analysis, Vol. 27, 1998, pp. 291-309.

Abstract: Imputation estimates based on imputed values obtained from neural network models used in an "impute first - aggregate next" approach, have been computed from Norwegian population census and administrative data. The imputation estimates were compared with simple unbiased estimates obtained by the traditional "aggreagate first - estimate next" approach and found superior for estimating proportions in small subgroups. predictors for predicting the accuracy of such imputation estimates were proposed. Results are promising for estimating small subgroup or area proportions.


Nordbotten, J.C. and Nordbotten, S. (1999): Search Patterns in Hypertext Exhibits.HICSS-32. Proceedings of The Thirty-Second Annual Hawaii International Conference on System Sciences. IEEE. 1999 ISBN 0-7695-0001-3. 7p.

Abstract: As access to hypermedia documents becomes generally available, it becomes increasingly important to understand how casual users search for information. We have studied search patterns used in a small hypermedia museum exhibit. Our goal has been to identify preferred search strategies that can be supported for effective information retrieval from web-based information systems. Our data indicates that users make little use of the embedded links enabling associative selection. Instead, topic selection was predominantly serially through first-in-index and next-page sequences. Further, sessions were short in both time and number of documents viewed, indicating little general interest in 'exploring' the exhibit.


Nordbotten, J.C. and Nordbotten, S. (1999): Search Patterns in Hypertext Exhibits.HICSS-32. Proceedings of The Thirty-Second Annual Hawaii International Conference on System Sciences. IEEE. 1999 ISBN 0-7695-0001-3. 7p.

Abstract:As access to electronic documents is made directly available to the general public from libraries, museums, schools, work places, and homes, it becomes increasingly important to understand how the casual user finds and accesses information within document collections. Museums are increasingly using hypertext documents to create virtual exhibits, either within the museum or on the World Wide Web. The hypertext format supports personalized, associative topic selection and is presumed to encourage user interest and information gathering.

We have studied usage patterns of a small hypertext exhibit presenting projects in the social sciences. Our goal has been to identify user search strategies that can be supported for effective information retrieval from web-based information systems.

The exhibit we've studied has been implemented on a stand-alone PC with touch screen input, at two locations. The 1st exhibit setup was available to the general public at the Bergen Museum, University of Bergen. The 2nd exhibit was placed in the main information area of the School of Social Science, University of Bergen. A web version, designed for usage testing, is scheduled for release in the summer of 1998.

To date, usage data has been collected for three populations. The 1st and 2nd samples were taken during the fall of 1996 and summer of 1997, when the primary museum visitors were students from the university and local schools and adult tourists, respectively. The 3rd sample was taken in the fall of 1997 after the exhibit was moved to the School of Social Science. Most users here had some connection to the social sciences and a presumed interest in the exhibit theme.

Our analysis of the data indicates that users made little use of the embedded links designed to support associative selection. Instead, topics were predominantly selected serially through first-in-index and next-page sequences. With the exception of a few users, sessions were short in both time and number of documents viewed, indicating little general interest in 'exploring' the exhibit.


Nordbotten, S. (2000): STATISTICS SWEDEN's EDITING PROCESS DATA PROJECTICES II. Proceedings of The Second International Conference on Establishment Surveys. Buffalo 2000. 7p.

Abstract:Statistics Sweden has continuously been working to improve the total quality of official statistics. One of the processes in which considerable work has been invested is in editing and imputation of statistical data. Recently a project for developing, evaluating and introducing methods for preserving process data has been established. The process data will be another dimension in the already formalized and implemented metadata system. One component will be the editing process data. This paper discusses some of the aspects of collecting, saving and using editing process data and outlines some possible approaches to the solution of this task.


Nordbotten, S. (2000):Evaluating Efficiency of Statistical Data Editing: General Framework ECE/UN Conference of European Statisticians, Methodological Material. GE-33259. 14 p.

Abstract: Data editing is a step in the preparation of statistics, the goal of which is to improve the quality of the statistical information. International research indicates that a typical statistical survey, the editing may consume up to 40% of all costs. The following questions have been reaised and are discussed in the paper:
- Is the use of these resources spent on editing justified?
- Can more effective editing strategies be appleied?
- Can the quality perhaps be improved by allocating some of the editing resources to prevent errors?


Nordbotten, S. (1999):Small Area Statistics from Survey and Imputed Data. Statistical Journal of the United nations ECE.Vol.14. pp. 297-309.

Abstract:This paper discusses improvement of statistics for small areas and groups utilizing sample survey data imputed by intensive use of background data fromm available census or administrative register sources. Reliable accuracy prediction for such statistics is importent and is also discussed and investigated in the paper.


Nordbotten, S.(1996): Meta-data about Editing and Accuracy for End Users. UN/ECE Work Session on Statistical Data Editing, Cardiff, UK, October 2000. UN/ECE Secretariat, Geneva.

Abstract:The paper focuses on end users of statistics and their needs for meata-data about the propducts and the production processes. The discussion is mainly limited to the quality dimension accuracy and the production process editing.


Copies of papers can be obtained by contacting the following adresses:

Return to publication pages.