Fuente: Wipo "IOT" The present invention relates to a method for performing adaptive machine learning in an Internet of things (IoT) environment, and more specifically, to a method for allowing various heterogeneous IoT devices operated in an IoT environment to adaptively perform distributed processing of machine learning according to each situation by recognizing an operating environment and a situation for a resource of a device. The method for performing adaptive machine learning in an IoT environment comprises the following steps of: (a) selecting any one IoT device among a plurality of IoT devices connected through networks by near field communications as a probing device; (b) allowing the selected probing device to transmit a resource information request message to the IoT devices; (c) allowing the IoT devices to transmit, to the probing device, a resource information response message in response to the resource information request message; (d) allowing the probing device to construct a resource tree for each of the IoT devices by using resource information included in the resource information response message received from each of the IoT devices; (e) allowing the probing device to determine a machine learning performance capability level for each of the IoT devices by using the resource tree; and (f) allowing the probing device to distribute an algorithm module of machine learning for an intellectual service provided in the IoT devices according to the machine learning performance capability level determined for each of the IoT devices. COPYRIGHT KIPO 2019