Hybrid seed selection and seed portfolio optimization by field
Techniques are provided for generating target success group of hybrid seeds for target fields include a server receiving agricultural data records that represent crop seed data describing seed and yield properties of hybrid seeds and first field geo-location data for agricultural fields where the hybrid seeds were planted. The server receives second geo-locations data for target fields where hybrid seeds are to be planted. The server generates a dataset of hybrid seed properties that include yield values and environmental classifications for hybrid seeds and then a dataset of success probability scores that describe the probability of a successful yield on the target fields based on the dataset of hybrid seed properties and the second geo-location data. The server generates target success yield group of hybrid seeds and probability of success values based on success probability scores and a yield threshold. The server causes display of the target success yield group.
1. A computer-implemented method comprising:
by an agricultural intelligence computer system, determining one or more agricultural data records that represent crop seed data describing seed and yield properties of one or more hybrid seeds and first field geo-location data for one or more current agricultural fields where the one or more hybrid seeds were planted;
by the agricultural intelligence computer system, determining second geo-location data for one or more target fields where hybrid seeds are to be planted;
by the agricultural intelligence computer system, generating a dataset of hybrid seed properties that describe a representative yield value for a particular growth cycle year, which represents a particular number of consecutive years a particular hybrid seed has been planted on a particular field, and an environmental classification for each hybrid seed of the one or more hybrid seeds from the one or more agricultural data records;
wherein the representative yield value is calculated as an average historical yield value from historical yield values representing the same particular growth cycle year observed from a set of agricultural fields;
by the agricultural intelligence computer system, generating a respective plurality of success probability scores, which describe a probability of a successful yield as a probability of success value on the one or more target fields, for each of the one or more hybrid seeds based upon the dataset of hybrid seed properties and the second geo-location data for the one or more target fields;
by the agricultural intelligence computer system, generating a target success yield group made up of a subset of the one or more hybrid seeds and a respective probability of success value associated with each of the subset of the one or more hybrid seeds that describes hybrid seeds that will produce a recommended yield estimate on the one or more target fields, based upon the plurality of success probability scores for each of the hybrid seeds and a configured probability threshold;
in response to the agricultural intelligence computer system determining that a difference between yield values of the one or more hybrid seeds and a calculated mean yield exceeds a threshold that comprises a calculated least significant difference value and a range of yield values that are within the calculated least significant difference value, by the agricultural intelligence computer system communicating with a controller for an agricultural machine, controlling the agricultural machine to plant, in the one or more target fields, the subset of the one or more hybrid seeds in the target success yield group.
2. The computer-implemented method of claim 1 , wherein field geo-location data includes observed relative maturity for the one or more hybrid seeds at the one or more agricultural fields.
3. The computer-implemented method of claim 1 , wherein the crop seed data includes at least one of: historical yield values, harvest time information, or relative maturity information for the one or more hybrid seeds at one or more agricultural fields.
4. The computer-implemented method of claim 1 , generating the dataset of hybrid seed properties that describe the representative yield value and the environmental classification for each hybrid seed of the one or more hybrid seeds further comprises calculating environmental classification values for each of the environmental classifications based upon observed relative maturity of the one or more hybrid seeds previously planted at the one or more agricultural.
5. The computer-implemented method of claim 4 , generating the plurality of success probability scores for each of the one or more hybrid seeds comprises performing logistic regression modelling on normalized yield values and the environmental classification values for each of the one or more hybrid seeds.
6. The computer-implement method of claim 4 , generating the plurality of success probability scores for each of the one or more hybrid seeds comprises performing random forest modelling on normalized yield values and the environmental classification values for each of the one or more hybrid seeds.
7. The computer-implement method of claim 4 , generating the plurality of success probability scores for each of the one or more hybrid seeds comprises performing support vector machine modelling on normalized yield values and the environmental classification values for each of the one or more hybrid seeds.
8. The computer-implement method of claim 4 , generating the plurality of success probability scores for each of the one or more hybrid seeds comprises performing gradient boosting machine modelling on normalized yield values and the environmental classification values for each of the one or more hybrid seeds.
9. The computer-implemented method of claim 1 , wherein the configured probability threshold is based on a configured yield value that is greater than a calculated range of average yield of hybrid seeds for the one or more target fields.
10. The computer-implemented method of claim 1 , displaying the target success yield group of the subset of the one or more hybrid seeds comprises:
sorting the subset of the one or more hybrid seeds by the probability of success value associated with each hybrid seed in descending order; and
displaying the sorted subset of the one or more hybrid seeds including the yield values.
11. The computer-implemented method of claim 1 , wherein said probability of success value on the one or more target fields comprises a probability of success value for the hybrid seed on the one or more target fields relative to others of the one or more hybrid seeds.
12. A server computer system comprising:
one or more processors;
one or more non-transitory computer-readable storage media storing instructions which, when executed using the one or more processors, cause the one or more processors to perform:
by an agricultural intelligence computer system, determining one or more agricultural data records that represent crop seed data describing seed and yield properties of one or more hybrid seeds and first field geo-location data for one or more current agricultural fields where the one or more hybrid seeds were planted;
by the agricultural intelligence computer system, determining second geo-location data for one or more target fields where hybrid seeds are to be planted;
by the agricultural intelligence computer system, generating a dataset of hybrid seed properties that describe a representative yield value for a particular growth cycle year, which represents a particular number of consecutive years a particular hybrid seed has been planted on a particular field, and an environmental classification for each hybrid seed of the one or more hybrid seeds from the one or more agricultural data records;
wherein the representative yield value is calculated as an average historical yield value from historical yield values representing the same particular growth cycle year observed from a set of agricultural fields;
by the agricultural intelligence computer system, generating a respective plurality of success probability scores, which describe a probability of a successful yield as a probability of success value on the one or more target fields, for each of the one or more hybrid seeds based upon the dataset of hybrid seed properties and the second geo-location data for the one or more target fields;
by the agricultural intelligence computer system, generating a target success yield group made up of a subset of the one or more hybrid seeds and a respective probability of success value associated with each of the subset of the one or more hybrid seeds that describes hybrid seeds that will produce a recommended yield estimate on the one or more target fields, based upon the plurality of success probability scores for each of the hybrid seeds and a configured probability threshold;
in response to the agricultural intelligence computer system determining that a difference between yield values of the one or more hybrid seeds and a calculated mean yield exceeds a threshold that comprises a calculated least significant difference value and a range of yield values that are within the calculated least significant difference value, by the agricultural intelligence computer system communicating with a controller for an agricultural machine, controlling the agricultural machine to plant, in the one or more target fields, the subset of the one or more hybrid seeds in the target success yield group.
13. The computer system of claim 12 , wherein field geo-location data includes observed relative maturity for the one or more hybrid seeds at the one or more agricultural fields.
14. The computer system of claim 12 , wherein the crop seed data includes at least one of: historical yield values, harvest time information, or relative maturity information for the one or more hybrid seeds at one or more agricultural fields.
15. The computer system of claim 12 , generating the dataset of hybrid seed properties that describe the representative yield value and the environmental classification for each hybrid seed of the one or more hybrid seeds further comprises calculating environmental classification values for each of the environmental classifications based upon observed relative maturity of the one or more hybrid seeds previously planted at the one or more agricultural.
16. The computer system of claim 15 , generating the plurality of success probability scores for each of the one or more hybrid seeds comprises performing linear regression modelling on normalized yield values and the environmental classification values for each of the one or more hybrid seeds.
17. The computer system of claim 15 , generating the plurality of success probability scores for each of the one or more hybrid seeds comprises performing random forest modelling on normalized yield values and the environmental classification values for each of the one or more hybrid seeds.
18. The computer system of claim 15 , generating the plurality of success probability scores for each of the one or more hybrid seeds comprises performing support vector machine modelling on normalized yield values and the environmental classification values for each of the one or more hybrid seeds.
19. The computer system of claim 15 , generating the plurality of success probability scores for each of the one or more hybrid seeds comprises performing gradient boosting machine modelling on normalized yield values and the environmental classification values for each of the one or more hybrid seeds.
20. The computer system of claim 12 , wherein the configured probability threshold is based on a configured yield value that is greater than a calculated range of average yield of hybrid seeds for the one or more target fields.
21. The computer system of claim 12 , displaying the target success yield group of the subset of the one or more hybrid seeds comprises:
sorting the subset of the one or more hybrid seeds by the probability of success value associated with each hybrid seed in descending order; and
displaying the sorted subset of the one or more hybrid seeds including the yield values.