SZ. BURGER ALICE: LATE ROMAN MONEY CIRCULATION IN SOUTH-PANNONIA / Régészeti Füzetek II/22. (Magyar Nemzeti Múzeum Budapest, 1981

THE METHODS OF THE ANALYSIS

I. THE METHODS OF THE ANALYSIS Any single coin can hold at about seven data item which are interesting from our point of wiev. I have tried to work out a method of evaluation a specific group of Roman coinage in order to be able to use computer aided methods in studying not only the dating facilities but the different relations concerning the above-mentioned data. The informations I have got with this methods yield a basis for further investigations of the history of economics of specific provinces in a given period. The investigations are build up on data collected in a time of ten years. The input data consists of published finding and of unpublished coins of my own works. The coinage of near twenty thousands copper coins of the late Roman times turned out to be a suitable amount of computer aided representative investigations. Based on the data about sites, Emperors and mints and divided by time intervalls, a great amount of new information has been found. Encouraged by it, Márkus András has used a clustering technique to enlighten six pairs of questions concerning the frequency distributions of the coinage. I should like to emphasise that the great amount of results has kept us from elaborating some more problems. By the analysis of the numerous types of revers it has come a temptation to analyse the types of the avers as well, but it has been impossible due to some couing problems related to the restricted resources available for us. The analysis of the avers by computer will be made at another occasion, but based on the data coded till now. The data stored till now gives the basis of further investigations on the numismatical characteristics of the different types of graves, or a detailed analysis of the mints Aquileia, Roma and Thessalonica, and some more problems. Other problems can be investigated on the basis of the data which are at your disposal as well. TRANSFORMING THE DATA TO THE INPUT OF THE COMPUTER The first task, namely the transformation of the card-index file to a coded input for the computer, turned to be a much bigger problem than I had thought. An unambigous, transparent coding system having only limited reguirements aqainst the coding efforts was to be formed, having realised that the decisions of this first phase (about omitting or combining data, forming groups of "others") sets the limits of all the later investigations. On the other hand the dimensions or grouping too fine which were not used later act as a burden on the restricted computing time and facilities. Having finished the work the second type of the problems turned out to be the more actual one. Due to the limited access to the computer some evaluations has been made by manual work, but based on the computer listings. The second step, the debugging of the data included no theoretical problems but some tinkering work. Thank for a program executing a lot of syntactical and semantical checks this step could be finished in a relatively short time. A compressed form of all the input data, we believe to be free of coding errors is available on magnetic tape containing more than 1100 card images, and on a listing easy to read. More data items may be included at any time. The coinage of the investigations has been divided into three groups following the phases: Phase A coins from 324 to 364 4 228 LRBC 3 Av/Rev only 620 608 Phase B coins from 364 to 375 RIC 13 Av/Rev only 334 596 Phase C corns from 375 to 378 RIC 106 The coins have been characterised by the following parameters: Emperor 8