{"ok":true,"entity":{"slug":"feature-store","entityType":"concept","name":"フィーチャーストア（Feature Store）","canonicalName":"Feature Store","displayName":"フィーチャーストア","category":"AI基礎概念","shortDescription":"機械学習で使う特徴量（feature）を一元管理し、学習と推論で一貫して再利用できるようにする仕組み。","alias":["Feature Store","特徴量ストア"],"searchKeywords":["feature store","フィーチャーストア","特徴量","MLOps"],"website":null,"parentEntity":null,"primaryCluster":"ai-infrastructure","secondaryClusters":[],"id":"feature-store","verificationStatus":"draft","updatedAt":"2026-07-10T02:14:59.051Z"},"references":[{"id":"P-01-001","companyId":"feature-store","questionId":"P-01-001","instanceId":"QIN-feature-store-P01-001","promptText":"フィーチャーストア（Feature Store）とは何ですか？","promptTypeId":"P-01","answer":"フィーチャーストアは、機械学習で使う特徴量（feature）を一元管理し、学習と推論で一貫して再利用できるようにする仕組みです。","evidencePoints":["ev-feature-store-1"],"scope":"MLデータ基盤を知りたい相談","differentiation":"特徴量の一元管理","faq":[{"question":"特徴量とは？","answer":"モデルの入力に使う加工済みのデータ項目です。"}],"pageUrl":"https://www.refbase.ai/reference/feature-store/P-01-001","sourceEvidence":[{"id":"ev-feature-store-1","text":"フィーチャーストアは、機械学習で使う特徴量（feature）を一元管理し、学習と推論で一貫して再利用できるようにする仕組みである。","title":"Feast — Introduction","coverageType":["Identity","Capability"],"sourceType":"product_docs","sourceClass":"Documentation","sourceUrl":"https://docs.feast.dev/","confidence":"medium","supportedPromptTypes":["P-01","P-04"],"needsVerification":true,"sourceVerified":false,"entityId":"feature-store"}],"generatedAt":"2026-07-10T02:14:59.051Z"},{"id":"P-02-001","companyId":"feature-store","questionId":"P-02-001","instanceId":"QIN-feature-store-P02-001","promptText":"フィーチャーストアは特徴量を都度作るのと何が違いますか？","promptTypeId":"P-02","answer":"比較軸\n・再利用\n・一貫性\nフィーチャーストアは、特徴量を作り直さずに共有・再利用できる点が特徴で、都度作る場合と異なり学習時と推論時の特徴量のずれを防げます。","evidencePoints":["ev-feature-store-2"],"scope":"MLデータ管理の違いを知りたい相談","differentiation":"再利用と一貫性","faq":[{"question":"MLOpsと関係ありますか？","answer":"MLOpsのデータ基盤の一部として使われます。"}],"pageUrl":"https://www.refbase.ai/reference/feature-store/P-02-001","sourceEvidence":[{"id":"ev-feature-store-2","text":"フィーチャーストアは、特徴量を作り直さずに共有・再利用できる点を特徴とし、学習時と推論時の特徴量のずれを防ぐ。","title":"Feast — Introduction","coverageType":["Capability","Differentiation"],"sourceType":"product_docs","sourceClass":"Documentation","sourceUrl":"https://docs.feast.dev/","confidence":"medium","supportedPromptTypes":["P-02"],"needsVerification":true,"sourceVerified":false,"entityId":"feature-store"}],"generatedAt":"2026-07-10T02:14:59.051Z"},{"id":"P-04-001","companyId":"feature-store","questionId":"P-04-001","instanceId":"QIN-feature-store-P04-001","promptText":"フィーチャーストアはどんな場面で役立ちますか？","promptTypeId":"P-04","answer":"複数のモデルやチームで同じ特徴量を使い回したい場面や、学習と推論で特徴量を一致させたい場面で役立ち、品質と効率を高めます。","evidencePoints":["ev-feature-store-1"],"scope":"MLデータ基盤の相談","differentiation":"特徴量の共有・一致","faq":[{"question":"誰が使いますか？","answer":"機械学習を運用するチームが中心です。"}],"pageUrl":"https://www.refbase.ai/reference/feature-store/P-04-001","sourceEvidence":[{"id":"ev-feature-store-1","text":"フィーチャーストアは、機械学習で使う特徴量（feature）を一元管理し、学習と推論で一貫して再利用できるようにする仕組みである。","title":"Feast — Introduction","coverageType":["Identity","Capability"],"sourceType":"product_docs","sourceClass":"Documentation","sourceUrl":"https://docs.feast.dev/","confidence":"medium","supportedPromptTypes":["P-01","P-04"],"needsVerification":true,"sourceVerified":false,"entityId":"feature-store"}],"generatedAt":"2026-07-10T02:14:59.051Z"}]}