FIELD: oil and gas.
SUBSTANCE: group of inventions relates to the field of searching for analogs of reservoirs with similar properties and filling in the missing values of descriptive attributes of the reservoir. A computer-implemented method for searching for analogs of deposits includes at least the following steps: get the first and second sample of records from at least one database of deposits and their attributes. Moreover, the first sample of records contains deposits of the first type, described by the first group of attributes, and the second sample of records contains deposits of the second type, described by the second group of attributes. The first sample to be marked out is formed; then a second one; the first and second sample to be marked out are marked with the help of at least two experts. Moreover, for each entry in a subgroup, experts mark whether this entry characterizes a deposit, which is an analogue of the target deposit in this subgroup. The first and second training samples are formed from the first and second labeled samples, respectively. The first classifier is trained using gradient boosting using the transformed attributes of the records of the first training sample; the second classifier is trained using gradient boosting using the transformed attributes of the records of the second training sample. Then the type of field is received from a user as well as the values of its attributes to determine its analogues; transforming the obtained attributes for use by the field classifier and searching for analogues using the trained classifier corresponding to the field type; then the user is presented with information about the search results.
EFFECT: increased efficiency of searching for oil reservoir analogs.
26 cl, 6 dwg
Title | Year | Author | Number |
---|---|---|---|
TRAINING CLASSIFIERS USED TO EXTRACT INFORMATION FROM NATURAL LANGUAGE TEXTS | 2018 |
|
RU2691855C1 |
CLASSIFIER TRAINING USED FOR EXTRACTING INFORMATION FROM TEXTS IN NATURAL LANGUAGE | 2018 |
|
RU2681356C1 |
SYSTEM AND METHOD FOR TWO-STAGE CLASSIFICATION OF FILES | 2018 |
|
RU2708356C1 |
METHOD FOR CHOOSING THE OPTIMAL HYDRAULIC FRACTURING DESIGN BASED ON THE INTELLIGENT ANALYSIS OF FIELD DATA TO INCREASE THE PRODUCTION OF HYDROCARBON RAW MATERIALS | 2021 |
|
RU2775034C1 |
METHOD OF CALCULATING CLIENT CREDIT RATING | 2019 |
|
RU2723448C1 |
RETRIEVAL OF INFORMATION OBJECTS USING A COMBINATION OF CLASSIFIERS ANALYZING LOCAL AND NON-LOCAL SIGNS | 2018 |
|
RU2686000C1 |
METHOD FOR DETERMINING PROBABILITY OF ACTH-ECTOPIC SYNDROME IN PATIENTS WITH ACTH-DEPENDENT ENDOGENOUS HYPERADRENOCORTICISM | 2023 |
|
RU2814146C1 |
METHOD FOR ATTRIBUTION OF PARTIALLY STRUCTURED TEXTS FOR FORMATION OF NORMATIVE-REFERENCE INFORMATION | 2020 |
|
RU2750852C1 |
FUEL COMBUSTION MODES MONITORING SYSTEM BY MEANS OF TORCH IMAGES ANALYSIS USING CLASSIFIER BASED ON CONVOLUTIONAL NEURAL NETWORK | 2018 |
|
RU2713850C1 |
SELECTION OF TEXT CLASSIFIER PARAMETER BASED ON SEMANTIC CHARACTERISTICS | 2016 |
|
RU2628431C1 |
Authors
Dates
2021-03-25—Published
2020-10-09—Filed