Hacktoberfest 2026：メンテナが10月に向けて印を付けた、オープンで初心者向けの issue。 Hacktoberfest の issue を見る

Incorrect lagged income quintiles after benefit-unit changes

オープン
  • 何か問題がありますか？
#543 コメント 0 件 リアクション 0 件 担当者 0 名 GitHub で見る

まだ誰も着手していません。

評価

難易度
4/5
見積もり時間
3〜5日
初心者へのやさしさ
45/100
issue の種類
バグ
明瞭さ
おおむね明確
活発さ
活発
技術スタック
java
領域
backend, databases

調査の方向性

Start with Person.getYHhQuintilesMonthC5L1() in src/main/java/simpaths/model/Person.java, the Q2–Q5 dummy getters, and new benefit unit initialization in src/main/java/simpaths/model/BenefitUnit.java. Trace how a person's benefit unit changes and decide how individual lagged quintile history should be retained, including how missing history is handled. Add focused checks for unchanged membership, a new unit at age 18, and partnership formation between people with different quintiles.

索引モデルが issue の本文から書いたものです。

説明

Coding problem

A person's lagged income quintile should describe the benefit unit they belonged to in the previous period. Instead, Person.getYHhQuintilesMonthC5L1() reads the lag stored on their current benefit unit:

return benefitUnit.getYHhQuintilesMonthC5L1();

When someone forms a new benefit unit, for example on turning 18 or forming a partnership, that unit may have no previous-period quintile. The Q2-Q5 dummy getters then all return zero, silently applying the Q1 reference-category coding. This is not evidence that the person was previously in Q1.

The new-unit initialisation supplies Q3 for a missing current quintile, but does not restore the person's previous quintile.

Verification

In a clean build of develop at 11bec2bd9 (23 October), an isolated getter test returned:

  • Original benefit unit with lagged Q5: Q5.
  • Membership changed to a new unit without history: null, with all Q2-Q5 dummy values 0.
  • Destination unit given lagged Q2: Q2, rather than the person's original Q5.

The same behaviour occurs in the saved income-validation JAR. These checks establish the getter problem; they do not measure its population frequency or aggregate effect on predictions.

Proposed coding repair

The approach discussed was to preserve each person's BU income quintile at the individual level before membership changes, and use that stored previous-period value for the lagged regressor. Two people forming a couple could then correctly retain different lagged quintiles. Retaining the previous BU link and its historical value is an alternative.

A focused test would check unchanged membership, a new BU at age 18, and partnership formation between people from different quintiles. Missing history would also need explicit handling rather than silently becoming the reference category. This repair preserves the meaning of the existing regressor; it does not itself require replacing it with parental income.

Separate education-specification question

Even with the coding repair, a student may subsequently have low own-BU income while still living with well-off parents. Whether E1a should instead use parental income, and how E1b should distinguish own from parental resources, is a separate specification decision requiring consistent estimation and Java implementation.

This issue is about preserving individual history, not choosing those education specifications or imputing parental income. It is also distinct from the equivalence-scale error in #541.

Could we confirm the preferred implementation and who will take it forward?

主要言語
Java
スター
26
フォーク
38
平均マージ
13時間 5分
マージ済み PR（30日）
7

環境構築

このプロジェクトには開発コンテナ、Dockerfile、コントリビューションガイドがありません。まず README を読み、一般的な手順ははじめてのコントリビューションガイドを参照してください。

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

simpaths/SimPaths のほかの issue

simpaths/SimPaths の issue をすべて見る

似ている issue

Java の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。