ASA flashcards
ASA · Build recall with a focused review session.
32 cards. Read the term and the definition here. The deck at the top of this page is optional.
- Unbiased estimator
- An estimator whose expected value equals the true parameter for every possible parameter value; bias is E[estimator] minus the parameter.
- Maximum likelihood estimator
- The parameter value that maximizes the likelihood function, or equivalently the log-likelihood, for the observed sample.
- Price elasticity of demand
- Percentage change in quantity demanded divided by percentage change in price; absolute values above one indicate elastic demand.
- Fisher equation
- Relates nominal interest rate, real interest rate, and expected inflation: nominal rate approximately equals real rate plus expected inflation.
- Accounting equation
- Assets equal liabilities plus equity; every transaction preserves this balance by affecting at least two accounts.
- Net present value
- Sum of projected cash flows discounted at a required rate; positive NPV means the project adds value.
- Bayes theorem
- Gives conditional probability P(A|B) as P(B|A)P(A)/P(B), combining prior and likelihood information.
- Central limit theorem
- For independent identically distributed variables with finite variance, the standardized sample mean converges to a standard normal distribution as sample size grows.
- Moment generating function
- M_X(t)=E[e^{tX}]; derivatives at zero yield moments, and it uniquely determines the distribution when it exists.
- Accumulation function
- a(t)=(1+i)^t for compound interest; it gives the future value at time t of one unit invested at time zero.
- Macaulay duration
- Weighted average time to cash flows, weights proportional to present values; measures interest rate sensitivity.
- Annuity-immediate
- Series of payments at end of each period; present value a_angle_n equals (1−v^n)/i.
- Pre-Actuarial Foundations Module
- SOA e-learning module covering foundational probability, statistics, financial mathematics, computing, and communication skills before advanced exams.
- Actuarial control cycle
- A framework: define the problem, develop a model, monitor results, and revise; used throughout actuarial practice.
- Force of mortality
- μ(x)=−S'(x)/S(x), the instantaneous death rate at age x given survival to x. It is also called the hazard rate.
- Net premium reserve
- Expected present value of future benefits minus expected present value of future net premiums, conditional on the policy being in force.
- Collective risk model
- Aggregate claims S=Σ X_i where claim count N and individual claim sizes X_i are random; often N is Poisson.
- Multiple decrement table
- Life table with several causes of decrement; dependent rates (aq)_x^{(j)} reflect competing risks. Each cause removes lives from the population.
- Credibility premium
- Weighted average Z times sample mean plus (1−Z) times prior mean; Z reflects sample size and variability.
- Lundberg bound
- For a risk process, ruin probability is at most e^{−Ru}, where R is the adjustment coefficient solving the Lundberg equation.
- Generalized linear model
- Extends linear regression by a link function g(E[Y])=Xβ and a specified exponential family distribution for the response.
- ARIMA model
- Time series model with autoregressive, differencing, and moving average components denoted ARIMA(p,d,q). The parameters specify lag orders and differencing degree.
- Principal components analysis
- Orthogonal linear transformations that maximize variance; components are ordered by eigenvalues and can reduce dimensionality.
- Actuarial Science Foundations Module
- SOA module integrating probability, financial mathematics, life contingencies, and risk concepts through applied modeling projects.
- Life table
- Table of l_x survivors, deaths d_x, and mortality rates q_x by age; used to compute insurance and annuity values.
- Bias-variance tradeoff
- Model error decomposes into bias, variance, and irreducible noise; increasing complexity lowers bias but raises variance.
- Random forest
- Ensemble of decision trees trained on bootstrap samples with random feature subsets; predictions average trees to reduce variance.
- Cross-validation
- Resampling method that splits data into folds to estimate out-of-sample predictive error and tune models.
- k-means clustering
- Unsupervised algorithm assigning observations to k clusters by minimizing within-cluster sum of squared distances to centroids.
- SHAP values
- Model-agnostic interpretability method based on Shapley values; explains each prediction's feature contributions. It attributes a prediction to individual features.
- Fundamentals of Actuarial Practice
- SOA e-learning course using case studies to teach actuarial practice, professionalism, communication, and enterprise risk management.
- Associateship Professionalism Course
- Required SOA course on professionalism, the Code of Professional Conduct, actuarial standards, and ethical judgment for Associateship.