Data storage system having means for compressing input data from sets of

Abstract

A data storage system utilizing a technique for compressing input data from cross-correlated sets of parameter values. The system dynamically allocates storage space based upon values of the input data. In systems where the multivalues parameters are grouped about a mean value, the data allocation techniques saves considerable memory area over a memory allocation for all possible values. A first embodiment uses one or more pointer matrix memories which store the addresses of blocks of counter locations. The first occurrence of a parameter set causes the assignment of a counter location for that set in a memory block whose address is stored in a pointer called by the occurrence of the set. Upon each subsequent occurrence of a set, the assigned counter location for the set is called through the pointer and incremented to maintain an accurate account of the frequency of the set occurrence. In an alternate embodiment, an encoding technique is utilized to embed address jump and occurrence count data in a single memory space. A data base is reduced by storing the count value of a particular parameter set in one memory location and associating with that location an address increment indicating the distance to the next parameter set. If the next parameter set is a predetermined distance away, then no address increment need be stored.