创建多选题生成器

0 浏览
0 试用
0 购买
Sep 29, 2025更新

根据指定主题生成五道多选题,内容精准且专业。

示例1

以下为“向量空间与基”主题的五道多选题(可多选),覆盖向量空间判定、子空间与基、维数与秩-零化度关系以及坐标表示的核心概念。每题均给出正确选项与简要解析,所据结论可由线性代数标准教材严格推出[1–3]。

题目一(向量空间判定,概念辨析,Bloom:理解/应用)
在实数域R上,以下哪些集合关于通常的加法与数乘构成向量空间(可多选)?
A. 所有次数不超过n的实系数多项式集合Pn
B. 所有次数恰为n的实系数多项式集合
C. 所有在区间[0,1]上的连续实值函数集合C[0,1]
D. 所有2×2可逆实矩阵的集合GL2(R)
E. 线性齐次方程组Ax=0在R^m中的全体解集合
正确答案:A, C, E
解析:A与C在通常运算下满足向量空间八公理[1];E是R^m的线性子空间(齐次解集,对加法与数乘封闭)[2]。B不含零向量且对加法不封闭;D对加法不封闭,也不含零矩阵[2]。

题目二(矩阵空间的基,Bloom:应用/分析)
令V=M2(R)为所有2×2实矩阵构成的向量空间,下列哪些集合是V的一组基(可多选)?记E11=[[1,0],[0,0]],E12=[[0,1],[0,0]],E21=[[0,0],[1,0]],E22=[[0,0],[0,1]],I为2×2单位阵。
A. {E11, E12, E21, E22}
B. {I, E12, E21, E22}
C. {E11, E12, E21, E11+E22}
D. {E11, E12, E21, 2E12}
E. {I, E11, E22, E12}
正确答案:A, B, C
解析:V维数为4[2]。A为标准基;B线性无关且生成V(E11=I−E22);C线性无关且生成V(E22=(E11+E22)−E11)。D线性相关(2E12与E12共线);E线性相关(I=E11+E22)[2]。

题目三(R3中的基判定,Bloom:应用)
以下哪些向量集合是R3的一组基(可多选)?
A. {(1,0,1),(0,1,1),(1,1,2)}
B. {(1,0,0),(0,1,0),(0,0,1)}
C. {(1,1,0),(1,0,1),(0,1,1)}
D. {(1,2,3),(2,4,6),(0,1,0)}
E. {(1,0,0),(0,1,0),(1,1,0)}
正确答案:B, C
解析:需线性无关且恰含三向量[1]。A中第三向量为前两者之和,相关;B为标准基;C行列式非零(|[ [1,1,0],[1,0,1],[0,1,1] ]|=-2),无关即为基;D前两向量共线,至多秩2;E第三向量落在前两者生成的平面内,相关[1]。

题目四(Pn的基与坐标,Bloom:应用/分析)
令P2为次数不超过2的实系数多项式空间。下列哪些集合是P2的一组基(可多选)?
A. {1, x, x^2}
B. {1, 1+x, 1+x+x^2}
C. {1+x, x+x^2, 1+x^2}
D. {1, x, 1+x}
E. {x, x^2, x^3}
正确答案:A, B, C
解析:P2维数为3[1]。A为标准基;B与C相对于{1,x,x^2}之坐标矩阵行列式非零,故线性无关并生成全空间;D线性相关(1+x=1+x);E不封闭于P2(x^3不在P2)[1]。

题目五(秩-零化度与行列空间,Bloom:分析/评价)
设A为实矩阵
A = [ [1, 2, 3],
      [0, 1, 1],
      [0, 0, 0] ]。
以下判断哪些正确(可多选)?
A. rank(A)=2
B. nullity(A)=1
C. A的前两列构成Col(A)的一组基
D. A的非零行向量构成Row(A)的一组基
E. Ax=0仅有零解
正确答案:A, B, C, D
解析:A为行阶梯形矩阵,主元列数为2,故秩为2;由秩-零化度定理,nullity(A)=3−2=1[2,3]。主元列在原矩阵中构成列空间的一组基[2];阶梯形矩阵的非零行向量构成行空间基[2]。E错误:零化度为1意味着存在非零解[3]。

技术增强的实施建议(针对在线测评与学习分析)
- 题库参数化与自动判分:在LMS(如Moodle、Canvas)中以随机参数生成选项矩阵与向量,结合题目五的矩阵结构自动计算秩与零化度(可用内嵌Python/NumPy或SageMathCell),实现个性化测评与抄袭抑制。
- 过程性反馈:在错选“非子空间”“线性相关”类干扰项时,触发基于子空间判别法(零向量、加法封闭、数乘封闭)与行列式/主元判定的分步提示[2]。
- 可视化与计算工具:引导学生用SymPy(.rref()、.rank()、.nullspace())或GeoGebra 3D直观检验线性无关与张成;配合Jupyter Notebook记录求解过程,支持格式化的学习证据留痕。
- 学习数据分析:利用选项级指标(如干扰项被选比例)定位共性误解(如将“可逆矩阵集合”误判为子空间),据此推送针对性微课与交互练习。

参考文献
[1] Axler, S. (2015). Linear Algebra Done Right (3rd ed.). Springer.
[2] Lay, D. C., Lay, S. R., & McDonald, J. J. (2016). Linear Algebra and Its Applications (5th ed.). Pearson.
[3] Strang, G. (2016). Introduction to Linear Algebra (5th ed.). Wellesley-Cambridge Press.

示例2

Below are five multiple-select questions (select all that apply) focused on process standards and digital tool design in technology-enhanced learning. Each item is followed by an answer key and concise rationale with citations.

Question 1: Governance of the Instructional Design Workflow
Which elements constitute a robust, standards-aligned workflow for developing digital learning resources in a higher-education setting?
A. Version control and change logs for learning materials (e.g., using a repository with documented approvals prior to release)
B. A Definition of Done that mandates accessibility peer review prior to publishing
C. Ad hoc SME sign-off after deployment to production
D. A requirements traceability matrix linking learning outcomes to usability/accessibility tests and acceptance criteria
E. Storing learner PII in unencrypted spreadsheets during development to speed iteration

Question 2: Interoperability Across a Multi-Tool Ecosystem
To ensure seamless launch, roster/grade return, and cross-platform analytics, which specifications should a design team prioritize?
A. 1EdTech LTI 1.3/Advantage for secure tool launch, deep linking, and grade services
B. xAPI for capturing learning experiences across web and mobile apps
C. IMS Caliper Analytics for event telemetry and common event vocabularies
D. Building only custom REST APIs with no adherence to sector standards
E. SCORM 1.2 as the primary mechanism for mobile offline tracking and deep linking

Question 3: Accessibility and Inclusive Design
Which design practices align with WCAG 2.2 AA and Universal Design for Learning (UDL) principles?
A. Provide captions and transcripts for prerecorded media
B. Maintain a minimum 4.5:1 color contrast ratio for body text
C. Lock navigation to a single linear path to ensure “focus”
D. Ensure full keyboard operability with visible focus indicators
E. Offer multiple means of representation (e.g., text alternatives for complex visuals)

Question 4: Data Protection and Ethical Analytics
Which practices align with GDPR/FERPA and privacy-by-design for learning analytics and remote proctoring?
A. Data minimization and purpose limitation in data collection
B. Conducting a Data Protection Impact Assessment (DPIA) for high-risk processing (e.g., biometric proctoring)
C. Storing pseudonymous learner identifiers with the re-identification key held separately
D. Indefinite retention of detailed clickstream data for “future research”
E. For biometric data, obtaining explicit consent and offering a non-biometric alternative to avoid coercion in a power-imbalanced educational context

Question 5: Usability and Quality in an Agile ID Process
Which actions reflect best practice for embedding usability and quality into iterative instructional design?
A. Formative usability testing with representative learners using think-aloud protocols during prototype sprints
B. Post-beta administration of standardized usability instruments (e.g., SUS or UMUX-Lite)
C. A/B testing of learning interventions with appropriate ethical review when generalizable knowledge is intended
D. Shipping to production first and relying solely on unstructured anecdotal feedback
E. Defining acceptance criteria using ISO/IEC 25010 quality characteristics (e.g., usability, reliability, security)

Answer Key
- Q1: A, B, D
- Q2: A, B, C
- Q3: A, B, D, E
- Q4: A, B, C, E
- Q5: A, B, C, E

Rationales (with references)
- Q1:
  - A and B: Formal change control and pre-release accessibility review align with human-centered, iterative quality practices (ISO 9241-210) and accessibility-by-design (W3C WCAG; Section 508).
  - D: Traceability supports verifiable alignment from requirements to tests and acceptance criteria (ISO/IEC 25010).
  - C and E are noncompliant; ad hoc sign-off lacks governance, and unencrypted PII violates security and privacy norms (ISO/IEC 27001; GDPR Art. 5, 25).

- Q2:
  - A: LTI 1.3/Advantage enables secure launch via OAuth 2.0/OpenID Connect, deep linking, NRPS, and grade services (1EdTech).
  - B and C: xAPI and Caliper support cross-platform learning event capture and common analytics vocabularies (ADL; 1EdTech).
  - D: Eschewing standards impairs interoperability and maintainability.
  - E: SCORM 1.2 does not support robust mobile offline tracking or deep linking and is constrained to LMS runtime contexts (ADL).

- Q3:
  - A, B, D, E: Captions/transcripts (WCAG 1.2.2), sufficient contrast (1.4.3), keyboard operability and focus visibility (2.1.1; 2.4.7), and multiple means of representation (UDL) constitute best practice.
  - C: Forced linear navigation can hinder learner agency and conflicts with UDL’s flexibility principles (CAST).

- Q4:
  - A: Data minimization and purpose limitation are core GDPR principles (Art. 5).
  - B: DPIA is required for high-risk processing like biometrics (GDPR Art. 35).
  - C: Pseudonymization and key separation reduce re-identification risk (GDPR Recital 28).
  - E: Biometric data are special-category; explicit consent must be freely given, which in education often requires non-coercive alternatives (GDPR Art. 9; EDPB Guidelines on consent).
  - D: Indefinite retention contravenes storage limitation (GDPR Art. 5(1)(e)); FERPA also restricts disclosure and emphasizes safeguarding education records.

- Q5:
  - A: Iterative formative testing is central to human-centered design (ISO 9241-210).
  - B: SUS and UMUX-Lite are validated instruments for summative usability assessment (Brooke; Lewis et al.).
  - C: Experimental comparisons require ethical oversight when intended to contribute to generalizable knowledge.
  - E: Using ISO/IEC 25010 helps specify verifiable quality attributes in acceptance criteria.
  - D: Post-hoc anecdotal feedback alone is insufficient and risks avoidable defects.

References
- 1EdTech Consortium. (n.d.). Learning Tools Interoperability (LTI) and Caliper Analytics specifications. https://www.1edtech.org
- ADL Initiative. (2017). Experience API (xAPI) specification. https://adlnet.gov/projects/xapi
- Brooke, J. (1996). SUS: A quick and dirty usability scale. In P. W. Jordan et al. (Eds.), Usability evaluation in industry. Taylor & Francis.
- CAST. (2018). Universal Design for Learning guidelines version 2.2. http://udlguidelines.cast.org
- European Data Protection Board. (2020). Guidelines 05/2020 on consent under Regulation 2016/679.
- European Union. (2016). General Data Protection Regulation (GDPR) 2016/679.
- ISO. (2019). ISO 9241-210:2019 Ergonomics of human-system interaction—Human-centred design for interactive systems.
- ISO/IEC. (2011). ISO/IEC 25010:2011 Systems and software engineering—SQuaRE—System and software quality models.
- ISO/IEC. (2022). ISO/IEC 27001:2022 Information security, cybersecurity and privacy protection—Information security management systems.
- Nielsen, J. (1994/2020). 10 usability heuristics for user interface design. Nielsen Norman Group.
- W3C. (2023). Web Content Accessibility Guidelines (WCAG) 2.2. https://www.w3.org/TR/WCAG22
- U.S. Department of Education. (n.d.). Family Educational Rights and Privacy Act (FERPA). https://www2.ed.gov/policy/gen/guid/fpco/ferpa/index.html

Note: Where version numbers are omitted (e.g., Caliper), the reference is to the current, maintained specification as published by the standards body at the time of consultation.

示例3

A continuación se presentan cinco ítems de selección múltiple (seleccione todas las opciones correctas) sobre fundamentos del aprendizaje supervisado. Cada ítem incluye opciones plausibles, la clave de respuesta y una justificación breve sustentada en la literatura especializada.

1) Naturaleza y objetivo del aprendizaje supervisado
Opciones:
A. Requiere un conjunto de entrenamiento con pares (x, y) etiquetados.
B. Su objetivo principal es maximizar la entropía de p(x).
C. Aprende una función f: X → Y que minimiza el riesgo empírico mediante una función de pérdida L.
D. No admite variables continuas como salida.
E. Puede formularse como la estimación de p(y|x) o de un mapeo determinista g(x).

Respuestas correctas: A, C, E.
Justificación: El aprendizaje supervisado utiliza ejemplos etiquetados para aprender un mapeo de entradas a salidas minimizando una función de pérdida (riesgo empírico) y puede adoptar enfoques discriminativos (p(y|x)) o funcionales directos (f(x)) (Bishop, 2006; Murphy, 2012; Goodfellow, Bengio, & Courville, 2016).

2) Partición de datos y validación
Opciones:
A. El escalado de características debe ajustarse únicamente con los datos de entrenamiento en cada partición o fold.
B. En series temporales, barajar aleatoriamente las observaciones antes de la validación mejora la estimación al romper la autocorrelación.
C. La estratificación en clasificación preserva la distribución de clases y mejora la estabilidad de las estimaciones.
D. Es apropiado usar el conjunto de prueba para seleccionar hiperparámetros si no se dispone de validación.
E. Con clases muy desbalanceadas, la validación estratificada y el uso de PR-AUC suelen ser más informativos que ROC-AUC.

Respuestas correctas: A, C, E.
Justificación: Para evitar leakage, el preprocesamiento se ajusta solo con entrenamiento en cada fold. La estratificación y métricas como PR-AUC son recomendables con clases raras. En series temporales deben usarse particiones respetuosas del orden temporal. El conjunto de prueba no debe guiar la selección de hiperparámetros (Kohavi, 1995; Saito & Rehmsmeier, 2015; Hastie, Tibshirani, & Friedman, 2009).

3) Funciones de pérdida y métricas de evaluación
Opciones:
A. MSE y MAE son pérdidas comunes en regresión.
B. La entropía cruzada binaria (log-loss) es adecuada para clasificación probabilística binaria.
C. La exactitud (accuracy) es siempre apropiada, aun con fuerte desbalance de clases.
D. ROC-AUC y PR-AUC capturan aspectos distintos; PR-AUC es más sensible a positivos raros.
E. En regresión, R² es una pérdida convexa recomendable para el entrenamiento.

Respuestas correctas: A, B, D.
Justificación: MSE/MAE son pérdidas estándar en regresión; la entropía cruzada es canónica en clasificación probabilística; con clases raras, PR-AUC resulta más informativa que ROC-AUC. Accuracy puede ser engañosa en escenarios desbalanceados. R² es una métrica, no una pérdida convexa típica de entrenamiento (Hastie et al., 2009; Murphy, 2012; Saito & Rehmsmeier, 2015).

4) Sobreajuste, sesgo-varianza y regularización
Opciones:
A. El sobreajuste se manifiesta con bajo error en entrenamiento y alto error en validación.
B. La regularización L2 tiende a fijar muchos coeficientes exactamente a cero, facilitando la selección de variables.
C. El early stopping funciona como un regularizador en procedimientos iterativos.
D. Aumentar la complejidad del modelo siempre reduce el error de generalización.
E. Aumentar el tamaño de datos o aplicar data augmentation puede reducir la varianza del modelo.

Respuestas correctas: A, C, E.
Justificación: El sobreajuste es discrepancia entre rendimiento en entrenamiento y generalización. L1, no L2, promueve soluciones exactamente esparsas. Early stopping y el aumento de datos actúan como regularizadores que mejoran la generalización (Geman, Bienenstock, & Doursat, 1992; Tibshirani, 1996; Goodfellow et al., 2016).

5) Algoritmos supervisados y sus propiedades
Opciones:
A. k-NN es un método supervisado basado en instancias; su coste de inferencia crece con el tamaño del conjunto de entrenamiento.
B. PCA es un método supervisado que maximiza la varianza de la variable objetivo y.
C. La regresión logística modela P(y=1|x) mediante una función sigmoide y es apropiada para clasificación binaria.
D. En SVM, un parámetro C más grande reduce la regularización y puede incrementar el sobreajuste.
E. Los árboles de decisión presentan alta varianza; técnicas de ensamble como bagging o Random Forest reducen dicha varianza.

Respuestas correctas: A, C, D, E.
Justificación: k-NN clasifica por proximidad y su inferencia depende del tamaño del entrenamiento; PCA es no supervisado; la regresión logística modela probabilidades condicionales; en SVM, un C grande penaliza fuertemente los errores de entrenamiento (menos regularización); los ensambles tipo bagging/Random Forest estabilizan árboles y reducen varianza (Hastie et al., 2009; Breiman, 2001; Bishop, 2006).

Referencias
- Bishop, C. M. (2006). Pattern Recognition and Machine Learning. Springer.
- Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32.
- Geman, S., Bienenstock, E., & Doursat, R. (1992). Neural networks and the bias/variance dilemma. Neural Computation, 4(1), 1–58.
- Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press.
- Hastie, T., Tibshirani, R., & Friedman, J. (2009). The Elements of Statistical Learning (2nd ed.). Springer.
- Kohavi, R. (1995). A study of cross-validation and bootstrap for accuracy estimation and model selection. In IJCAI (pp. 1137–1145).
- Murphy, K. P. (2012). Machine Learning: A Probabilistic Perspective. MIT Press.
- Saito, T., & Rehmsmeier, M. (2015). The precision–recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS ONE, 10(3), e0118432.
- Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B, 58(1), 267–288.

适用用户

高校教师与教务

为课程章节快速生成契合教学目标的多选题,用于课前诊断、课中测验与课后巩固,减少命题与审题工作量。

企业培训经理(HRD)

围绕数字工具与流程规范,一键产出岗位能力测评题组,用于新人训练营、认证考试和年度复训评估。

在线课程制作人(MOOC教研)

批量构建课程题库,自动匹配学术表达与多语言版本,便捷上线国内外平台并提升完课率。

教学设计师(Instructional Designer)

在设计混合式与项目式学习方案时,快速生成聚焦技术选型与教学策略的测评题,验证目标达成度。

教育科技产品经理

用于用户教育与功能引导,搭建标准化测评题集,收集学习数据,支持版本迭代与增长活动。

教研机构与考试中心

建立高质量试题样板,统一命题规范与难度梯度,缩短审题周期,支撑证书考试与等级测评。

解决的问题

用一条可复用的高阶提示词,帮助教师、教研员、培训经理与题库运营,在最短时间内围绕指定主题生成5道高质量多选题,兼顾专业深度与可测性,直接服务于单元测验、在线课程闯关、认证考前演练与企业培训抽测等场景。 - 保证专业与准确:以“教育技术专家”视角命题,紧扣核心概念与应用场景,减少歧义与表述漏洞。 - 难度与干扰项更科学:题干逻辑严谨,干扰项具备迷惑度与区分度,避免套路化与常识题。 - 多语言与学术规范:支持指定作答语言(如中文/英文),整体表达遵循学术写作规范;当题目涉及事实性陈述时,提供可信来源线索,提升权威性。 - 输出结构清晰:一键获得可直接投放的多选题清单,减少二次编辑成本。 - 业务价值:降低命题人力与沟通成本,标准化评测质量,显著缩短从选题到上线的周期,帮助你更快验证学习效果并持续迭代课程。

特征总结

按主题一键生成五道多选题,覆盖关键概念,题干精准,直连教学目标与评估需求。
面向教育技术场景出题,融入在线学习、数字工具与教学设计情境,贴合真实教学与培训。
自动遵循学术写作风格,表达严谨有据,便于直接纳入高校课程与企业认证考试。
支持多语言出题与呈现,一键切换语言版本,轻松覆盖跨地区培训与国际课程。
可按给定场景定制题目侧重,强化技术选型、教学策略与实施能力的诊断与评估。
自动规避离题与冗余,聚焦核心能力考察,减少来回修改与审题时间成本。
输出结构清晰,含论点与依据表达,方便后续补充解析与资料,快速形成合格题库。
注重事实核查与可信来源引用,降低知识性错误风险,提升测评结果的公信力与说服力。
模板可复用并轻松扩展到不同课程模块,统一命题口径与质量,支持批量化备考素材.

如何使用购买的提示词模板

1. 直接在外部 Chat 应用中使用

将模板生成的提示词复制粘贴到您常用的 Chat 应用(如 ChatGPT、Claude 等),即可直接对话使用,无需额外开发。适合个人快速体验和轻量使用场景。

2. 发布为 API 接口调用

把提示词模板转化为 API,您的程序可任意修改模板参数,通过接口直接调用,轻松实现自动化与批量处理。适合开发者集成与业务系统嵌入。

3. 在 MCP Client 中配置使用

在 MCP client 中配置对应的 server 地址,让您的 AI 应用自动调用提示词模板。适合高级用户和团队协作,让提示词在不同 AI 工具间无缝衔接。

¥20.00元
平台提供免费试用机制,
确保效果符合预期,再付费购买!

您购买后可以获得什么

获得完整提示词模板
- 共 236 tokens
- 2 个可调节参数
{ 输入主题 } { 输出语言 }
自动加入"我的提示词库"
- 获得提示词优化器支持
- 版本化管理支持
获得社区共享的应用案例
限时免费

不要错过!

免费获取高级提示词-优惠即将到期

17
:
23
小时
:
59
分钟
:
59