公开(公告)号：US10817657B2

公开(公告)日：2020-10-27

申请号：US16232279

申请日：2018-12-26

Applicant: Nokia Solutions and Networks Oy

Inventor： Ahmet Akyamac ,  Rajarshi Bhowmik ,  Jeongran Lee

IPC: G06F40/177 ,  G06F40/174 ,  G06N5/04 ,  G06F16/35

Abstract: Various example embodiments for supporting determination of field types in tabular data are presented. Various example embodiments for supporting determination of field types in tabular data are configured to provide improvements in computer performance for supporting determination of field types in tabular data. Various example embodiments for supporting determination of field types in tabular data are configured to generate an embedding knowledge base based on training tabular data and to process new tabular data based on the embedding knowledge base in order to determine field types of fields included in the new tabular data.

公开(公告)号：US11556551B2

公开(公告)日：2023-01-17

申请号：US17103467

申请日：2020-11-24

Applicant: Nokia Solutions and Networks Oy

Inventor： Ahmet Akyamac ,  Huseyin Uzunalioglu

IPC: G06F17/00 ,  G06F16/2458 ,  G06N20/00

Abstract: Systems, methods, and software of processing a tabular dataset. In one embodiment, a system extracts raw association rules from the tabular dataset. Each of the raw association rules comprises a relationship between a set of antecedents and a single consequent, and corresponds to one or more transactions. The system determines potential rule merge groups of the raw association rules based on the antecedents, and determines one or more actual rule merge groups of the raw association rules in each potential rule merge group based on the transactions. The system combines the raw association rules in an actual rule merge group to generate a merged association rule. The system then generates a set of insights based on one or more merged association rules, and performs an operation based on the set of insights.

公开(公告)号：US20220164347A1

公开(公告)日：2022-05-26

申请号：US17103467

申请日：2020-11-24

Applicant: Nokia Solutions and Networks Oy

Inventor： Ahmet Akyamac ,  Huseyin Uzunalioglu

IPC: G06F16/2458 ,  G06N20/00

Abstract: Systems, methods, and software of processing a tabular dataset. In one embodiment, a system extracts raw association rules from the tabular dataset. Each of the raw association rules comprises a relationship between a set of antecedents and a single consequent, and corresponds to one or more transactions. The system determines potential rule merge groups of the raw association rules based on the antecedents, and determines one or more actual rule merge groups of the raw association rules in each potential rule merge group based on the transactions. The system combines the raw association rules in an actual rule merge group to generate a merged association rule. The system then generates a set of insights based on one or more merged association rules, and performs an operation based on the set of insights.

公开(公告)号：US20200210520A1

公开(公告)日：2020-07-02

申请号：US16232279

申请日：2018-12-26

Applicant: Nokia Solutions and Networks Oy

Inventor： Ahmet Akyamac ,  Rajarshi Bhowmik ,  Jeongran Lee

IPC: G06F17/24 ,  G06F16/35 ,  G06N5/04

Abstract: Various example embodiments for supporting determination of field types in tabular data are presented. Various example embodiments for supporting determination of field types in tabular data are configured to provide improvements in computer performance for supporting determination of field types in tabular data. Various example embodiments for supporting determination of field types in tabular data are configured to generate an embedding knowledge base based on training tabular data and to process new tabular data based on the embedding knowledge base in order to determine field types of fields included in the new tabular data.