Abstract: The present invention provides a data analysis system capable of performing analysis appropriately using a CNN model while reducing communication traffic. A data analysis system 90 includes: an instrument 10 that performs a conversion process of outputting compression data obtained as a result of processing observation data received via an input layer of a learned neural network 18A using portions ranging from the input layer to a predetermined intermediate layer; and a device 20 that performs an analysis process of inputting the compression data to a subsequent intermediate layer in a learned neural network 18B, inputting data obtained by decoding the compression data, which is an output of the subsequent intermediate layer, to an output layer configured using a CNN model, and obtaining an analysis result of the observation data as an output of the output layer.

Abstract: A porous multi-layer film favorably employable as a separator film for a nonaqueous electric battery comprises at least three united polyolefin layers, in which at least one polyolefin layer is a polyethylene layer and at least one polyolefin layer is a polypropylene layer which is placed in contact with the polyethylene layer. The polyolefin layers are combined to form a united structure with a peel strength of at least 3 g/15 mm and has a pore volume of 30-80%, a maximum pore size of 0.02-2 .mu.m, a shutdown temperature of 135.degree.-140.degree. C., and a thermal durability to maintain the shutdown condition up to at least 180.degree. C.

Abstract: A process for producing a microporous stretched polyolefin film having a large number of micropores, whereby an unstretched polyolefin film is stretched to predetermined strain amount at a constant stretching temperature by using different strain rates in the initial stage to the final stage. A microporous polyolefin film obtained by the above process can retain a film state even at the melting point or higher temperatures when the film is fixed at the periphery.