US 12,422,821 B2
Method of setting factor variable area, and system
Mitsunori Kajima, Matsumoto (JP)
Assigned to SEIKO EPSON CORPORATION, Tokyo (JP)
Filed by SEIKO EPSON CORPORATION, Tokyo (JP)
Filed on Aug. 5, 2022, as Appl. No. 17/817,722.
Claims priority of application No. 2021-130455 (JP), filed on Aug. 10, 2021.
Prior Publication US 2023/0047908 A1, Feb. 16, 2023
Int. Cl. G05B 19/4155 (2006.01)
CPC G05B 19/4155 (2013.01) [G05B 2219/31372 (2013.01)] 4 Claims
OG exemplary drawing
 
1. A method of setting a factor variable area, the method comprising:
retrieving from a memory
a plurality of measured values of at least one factor variable, and
a label indicating good quality or bad quality of a product or a manufacturing process of producing the product corresponding to each of the plurality of measured values;
dividing a factor variable space, defined by the at least one factor variable, into a plurality of grids by equally dividing a range determined by a maximum value and a minimum value of the plurality of measured values for each factor variable of the at least one factor variable;
setting a plurality of candidate areas based on the division, wherein
each of the plurality of candidate areas includes one grid of the plurality of grids or a plurality of adjacent grids of the plurality of grids;
deriving, for the each of the plurality of candidate areas, a density of each grid of the plurality of grids based on the label associated with a measured value of the plurality of measured values that is within the each of the plurality of candidate areas;
acquiring, as a dummy value of the at least one factor variable, a value of a position where each grid of the plurality of grids is equally divided for the each factor variable of the at least one factor variable;
deriving an estimated label indicating estimated good quality or estimated bad quality of the product or the manufacturing process of producing the product by inputting the dummy value to a determination model that learned a correspondence relationship between the measured value and the label;
deriving, for the each of the plurality of candidate areas, a first density and a second density of the each grid of the plurality of grids, wherein
the first density is based on the label associated with the measured value, and
the second density is based on the estimated label associated with the dummy value that are within a candidate area of the plurality of candidate areas; and
selecting one of the plurality of candidate areas as the factor variable area, based on the first density and the second density, wherein the factor variable area defines a range of a value of the at least one factor variable based on representation of a condition of the manufacturing process as the value of the at least one factor variable.