Abstract: An approach is provided for rationalizing operational data. Persisted knowledge of previous data transactions is determined to not include a previous data profile matching a current data profile. Stochastic data sampling generates a sample of a current dataset. Data size and shape corrections are determined and updated by iterations of a statistical inference technique. Data in the sample is corrected by using the data size and shape corrections to provide a confidence level exceeding a threshold. Persisted knowledge configurations are updated with parameters of the statistical inference technique, the data size and shape corrections, and filtering rule(s). A recommended modification of a query of a current data transaction is determined to provide data in the current dataset with the level of confidence exceeding the threshold. The recommended modification is applied to the query.

Abstract: An approach is provided for rationalizing operational data. Persisted knowledge of previous data transactions is determined to not include a previous data profile matching a current data profile. Stochastic data sampling generates a sample of a current dataset. Data size and shape corrections are determined and updated by iterations of a statistical inference technique. Data in the sample is corrected by using the data size and shape corrections to provide a confidence level exceeding a threshold. Persisted knowledge configurations are updated with parameters of the statistical inference technique, the data size and shape corrections, and filtering rule(s). A recommended modification of a query of a current data transaction is determined to provide data in the current dataset with the level of confidence exceeding the threshold. The recommended modification is applied to the query.

Abstract: An approach is provided for rationalizing operational data. A current data profile of a current dataset utilized by a current data transaction is determined. Persisted knowledge of previous data transactions is determined to include a previous data profile of a previous dataset that matches the current data profile. If the persisted knowledge indicates that data size and shape corrections were applied to the previous dataset, filtering corrections of the current dataset based on the data size and shape corrections are determined, the persisted knowledge is rationalized based on the filtering corrections, and queries of the current transaction are modified based on the filtering corrections, or if the data size and shape corrections were not applied to the previous dataset, the persisted knowledge is rationalized based on the data profile match and the queries of the current transaction are modified based on the rationalized persisted knowledge.

Abstract: An approach is provided for rationalizing operational data. A current data profile of a current dataset utilized by a current data transaction is determined. Persisted knowledge of previous data transactions is determined to include a previous data profile of a previous dataset that matches the current data profile. If the persisted knowledge indicates that data size and shape corrections were applied to the previous dataset, filtering corrections of the current dataset based on the data size and shape corrections are determined, the persisted knowledge is rationalized based on the filtering corrections, and queries of the current transaction are modified based on the filtering corrections, or if the data size and shape corrections were not applied to the previous dataset, the persisted knowledge is rationalized based on the data profile match and the queries of the current transaction are modified based on the rationalized persisted knowledge.

Abstract: An approach is presented for managing resources of a field-programmable gate array (FPGA). At runtime, first data is extracted and processed. At runtime and via a very high speed integrated circuit hardware description language (VHDL) interface, a change in a size, a structure, or a load schedule of next data is received. The change is determined by a quantitative method analyzing the first data and executing external to the FPGA. At runtime, a first bootstrap code in the FPGA executes, and in response, other bootstrap codes in the FPGA are updated. The first bootstrap code is configured to update the structure of the next data. The other bootstrap codes are configured to extract and process, and to determine an order of processing and a configuration of the next data. The next data is extracted and processed based on the updated other bootstrap codes.

Abstract: An approach is presented for managing resources of a field-programmable gate array (FPGA). At runtime, first data is extracted and processed. At runtime and via a very high speed integrated circuit hardware description language (VHDL) interface, a change in a size, a structure, or a load schedule of next data is received. The change is determined by a quantitative method analyzing the first data and executing external to the FPGA. At runtime, a first bootstrap code in the FPGA executes, and in response, other bootstrap codes in the FPGA are updated. The first bootstrap code is configured to update the structure of the next data. The other bootstrap codes are configured to extract and process, and to determine an order of processing and a configuration of the next data. The next data is extracted and processed based on the updated other bootstrap codes.