ARBIMETRA LAB
Model Explorer
Search 78 statistical, econometric, financial and machine-learning models. Each one states its data needs and assumptions, and shows how it was validated.
78 models
Descriptive Statistics & Data Analysis
-
Model 1Validated
Descriptive Statistics
Summarises the centre, spread and size of numeric variables, keeping sample and population formulas separate.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- One or more numeric variables
- Data
- Numeric
- Assumes
- Values are measured on a numeric scale
- Level
- Beginner
-
Model 2Validated
Frequency Distribution
Counts how many observations fall in each class or category, with relative and cumulative frequencies.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- One variable
- Data
- Numeric, Categorical
- Assumes
- Classes do not overlap and cover all values
- Level
- Beginner
-
Model 3Validated
Percentile & Quartile Analysis
Finds the values below which given shares of the data lie, along with quartiles, the interquartile range and percentile ranks.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- One numeric variable
- Data
- Numeric
- Assumes
- Values can be ordered
- Level
- Beginner
-
Model 4Validated
Skewness & Kurtosis
Measures asymmetry and tail weight, reporting sample-adjusted, population and Pearson versions.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- One or more numeric variables
- Data
- Numeric
- Assumes
- At least 4 valid values
- Level
- Intermediate
-
Model 5Validated
Correlation Analysis
Measures linear and rank association between pairs of variables, with p-values, intervals, a matrix and a heatmap.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- Two or more numeric variables
- Data
- Numeric
- Assumes
- Pearson: a linear relationship
- Level
- Beginner
-
Model 6Validated
Covariance Analysis
Shows how variables move together in their original units, with sample and population matrices.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- Two or more numeric variables
- Data
- Numeric
- Assumes
- Observations are paired
- Level
- Beginner
-
Model 7Validated
Cross-tabulation
Builds a two-way table with row, column and total percentages, and an optional chi-square test.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- Two categorical variables
- Data
- Categorical
- Assumes
- Each row of data is one independent observation
- Level
- Beginner
-
Model 8Validated
Missing Value Analysis
Reports missing values by variable, complete and incomplete rows and common missingness patterns.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- A dataset
- Data
- Any
- Assumes
- None for counting
- Level
- Beginner
-
Model 9Validated
Outlier Detection
Flags unusual values using the IQR rule, z-scores or the robust modified z-score based on the median absolute deviation.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- One numeric variable
- Data
- Numeric
- Assumes
- z-scores work best for roughly symmetric data
- Level
- Beginner
-
Model 10Validated
Distribution Analysis
Compares the shape of a variable with a normal distribution using a histogram, kernel density, box plot, Q-Q plot and empirical CDF.
- Category
- Descriptive Statistics & Data Analysis
- Needs
- One numeric variable
- Data
- Numeric
- Assumes
- At least 8 valid values
- Level
- Intermediate
Inferential Statistics & Hypothesis Testing
-
Model 11Validated
One-sample T-test
Tests whether a population mean differs from a specified value.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- One numeric variable; Reference mean
- Data
- Numeric
- Assumes
- Observations are independent
- Level
- Beginner
-
Model 12Validated
Independent Samples T-test
Compares the means of two independent groups, with both pooled-variance and Welch versions.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- Numeric outcome; Two groups
- Data
- Numeric, Categorical
- Assumes
- The groups are independent
- Level
- Beginner
-
Model 13Validated
Paired Samples T-test
Tests whether the mean difference between paired measurements is zero.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- Two numeric columns with matched rows
- Data
- Numeric
- Assumes
- Pairs are independent of each other
- Level
- Beginner
-
Model 14Validated
Welch's T-test
Compares two independent means without assuming equal variances, using Welch-Satterthwaite degrees of freedom.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- Numeric outcome; Two groups
- Data
- Numeric, Categorical
- Assumes
- The groups are independent
- Level
- Beginner
-
Model 15Validated
One-way ANOVA
Tests whether several group means are equal, with Tukey HSD comparisons for pairs.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- Numeric outcome; Three or more groups
- Data
- Numeric, Categorical
- Assumes
- Observations are independent
- Level
- Intermediate
-
Model 16Validated
Two-way ANOVA
Tests the main effect of two factors and their interaction.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- Numeric outcome; Two categorical factors
- Data
- Numeric, Categorical
- Assumes
- Observations are independent
- Level
- Intermediate
-
Model 17Validated
Chi-square Test
Tests whether categories follow an expected pattern, or whether two categorical variables are associated.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- One or two categorical variables
- Data
- Categorical
- Assumes
- Observations are independent
- Level
- Beginner
-
Model 18Validated
Mann-Whitney U Test
Compares two independent groups using ranks, without assuming normality.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- Outcome; Two groups
- Data
- Numeric, Categorical
- Assumes
- Observations are independent
- Level
- Intermediate
-
Model 19Validated
Wilcoxon Signed-rank Test
Tests whether paired differences are centred at zero, using signed ranks.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- Two numeric columns with matched rows
- Data
- Numeric
- Assumes
- Differences are symmetric around their centre
- Level
- Intermediate
-
Model 20Validated
Kruskal-Wallis Test
Compares several independent groups using ranks, with Dunn post-hoc comparisons.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- Outcome; Three or more groups
- Data
- Numeric, Categorical
- Assumes
- Observations are independent
- Level
- Intermediate
-
Model 21Validated
Shapiro-Wilk Test
Tests whether a sample could have come from a normal distribution.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- One numeric variable
- Data
- Numeric
- Assumes
- Observations are independent
- Level
- Beginner
-
Model 22Validated
Levene's Test
Tests whether groups have equal variances, centred on the mean, median or trimmed mean.
- Category
- Inferential Statistics & Hypothesis Testing
- Needs
- Outcome; Two or more groups
- Data
- Numeric, Categorical
- Assumes
- Observations are independent
- Level
- Intermediate
Regression & Financial Econometrics
-
Model 23Validated
Simple Linear Regression
Estimates a straight-line relationship between two variables with inference, a fitted line and diagnostics.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; One independent variable
- Data
- Numeric
- Assumes
- Linear in coefficients
- Level
- Beginner
-
Model 24Validated
Multiple Linear Regression (OLS)
Estimates a linear model with several explanatory variables, an ANOVA table, diagnostics and predictions.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; One or more independent variables
- Data
- Numeric
- Assumes
- Linear in coefficients
- Level
- Intermediate
-
Model 25Validated
Polynomial Regression
Fits a curve using powers of one variable, still linear in its coefficients.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; One explanatory variable
- Data
- Numeric
- Assumes
- Linear in coefficients
- Level
- Intermediate
-
Model 26Validated
Weighted Least Squares
Estimates a regression where each observation has its own weight.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; Independent variables; Positive weights
- Data
- Numeric
- Assumes
- Weights are proportional to the inverse of the error variance
- Level
- Intermediate
-
Model 27Validated
Logistic Regression
Models the probability of a binary outcome, with odds ratios, marginal effects and classification measures.
- Category
- Regression & Financial Econometrics
- Needs
- Binary dependent variable; Independent variables
- Data
- Numeric, Binary
- Assumes
- Observations are independent
- Level
- Intermediate
-
Model 28Validated
Probit Regression
Models a binary outcome through the standard normal distribution, with average marginal effects.
- Category
- Regression & Financial Econometrics
- Needs
- Binary dependent variable; Independent variables
- Data
- Numeric, Binary
- Assumes
- Observations are independent
- Level
- Intermediate
-
Model 29Validated
Ridge Regression
Shrinks coefficients with an L2 penalty to stabilise estimates when predictors are correlated.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; Independent variables
- Data
- Numeric
- Assumes
- Training and test rows come from the same population
- Level
- Advanced
-
Model 30Validated
LASSO Regression
Applies an L1 penalty that can set coefficients exactly to zero.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; Independent variables
- Data
- Numeric
- Assumes
- Training and test rows come from the same population
- Level
- Advanced
-
Model 31Validated
Elastic Net
Combines L1 and L2 penalties to select variables while handling correlated predictors.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; Independent variables
- Data
- Numeric
- Assumes
- Training and test rows come from the same population
- Level
- Advanced
-
Model 32Validated
Quantile Regression
Estimates how a chosen conditional quantile, such as the median, depends on the variables.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; Independent variables
- Data
- Numeric
- Assumes
- The conditional quantile is linear in the variables
- Level
- Advanced
-
Model 33Validated
Instrumental Variables / 2SLS
Estimates a regression with endogenous regressors using instruments, with relevance, overidentification and endogeneity diagnostics.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; Endogenous variable(s); Instrument(s)
- Data
- Numeric
- Assumes
- Instruments are relevant
- Level
- Advanced
-
Model 34Validated
Panel Fixed Effects
Controls for stable differences between entities using the within transformation, with optional time effects.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; Independent variables; Entity identifier
- Data
- Panel
- Assumes
- Regressors are strictly exogenous given the entity effect
- Level
- Advanced
-
Model 35Validated
Panel Random Effects
Estimates a panel model using Swamy-Arora variance components, with a Hausman comparison to fixed effects.
- Category
- Regression & Financial Econometrics
- Needs
- Dependent variable; Independent variables; Entity identifier
- Data
- Panel
- Assumes
- Entity effects are uncorrelated with all regressors
- Level
- Advanced
Time Series & Financial Forecasting
-
Model 36Validated
Time Series Decomposition
Splits a series into trend, seasonal and residual components, additively or multiplicatively.
- Category
- Time Series & Financial Forecasting
- Needs
- One series; Seasonal period
- Data
- Time series
- Assumes
- The seasonal pattern repeats
- Level
- Intermediate
-
Model 37Validated
Moving Average
Smooths a series with a simple or linearly weighted trailing average.
- Category
- Time Series & Financial Forecasting
- Needs
- One series; Window
- Data
- Time series
- Assumes
- Rows are in time order
- Level
- Beginner
-
Model 38Validated
Augmented Dickey-Fuller Test
Tests whether a series has a unit root, with a choice of deterministic terms and lag selection.
- Category
- Time Series & Financial Forecasting
- Needs
- One series
- Data
- Time series
- Assumes
- Rows are in time order
- Level
- Intermediate
-
Model 39Validated
Phillips-Perron Test
Tests for a unit root, correcting for serial correlation with a long-run variance rather than extra lags.
- Category
- Time Series & Financial Forecasting
- Needs
- One series
- Data
- Time series
- Assumes
- Rows are in time order
- Level
- Intermediate
-
Model 40Validated
KPSS Test
Tests level or trend stationarity as the null hypothesis, which complements unit-root tests.
- Category
- Time Series & Financial Forecasting
- Needs
- One series
- Data
- Time series
- Assumes
- Rows are in time order
- Level
- Intermediate
-
Model 41Validated
Autoregressive Model (AR)
Models a series using its own past values.
- Category
- Time Series & Financial Forecasting
- Needs
- One series; Order p
- Data
- Time series
- Assumes
- The series is stationary after any differencing
- Level
- Intermediate
-
Model 42Validated
Moving Average Model (MA)
Models a series using past forecast errors.
- Category
- Time Series & Financial Forecasting
- Needs
- One series; Order q
- Data
- Time series
- Assumes
- The series is stationary after any differencing
- Level
- Intermediate
-
Model 43Validated
ARMA
Combines autoregressive and moving-average terms for a stationary series.
- Category
- Time Series & Financial Forecasting
- Needs
- One series; Orders p and q
- Data
- Time series
- Assumes
- The series is stationary after any differencing
- Level
- Intermediate
-
Model 44Validated
ARIMA
Models a differenced series with AR and MA terms, with diagnostics, forecasts, intervals, AIC and BIC.
- Category
- Time Series & Financial Forecasting
- Needs
- One series; Orders p, d and q
- Data
- Time series
- Assumes
- The series is stationary after any differencing
- Level
- Intermediate
-
Model 45Validated
SARIMA
Extends ARIMA with seasonal terms for series with repeating patterns.
- Category
- Time Series & Financial Forecasting
- Needs
- One series; Orders p, d, q, P, D, Q and s
- Data
- Time series
- Assumes
- The series is stationary after any differencing
- Level
- Advanced
-
Model 46Validated
Exponential Smoothing
Forecasts with simple exponential smoothing, Holt's trend method or Holt-Winters seasonal methods.
- Category
- Time Series & Financial Forecasting
- Needs
- One series
- Data
- Time series
- Assumes
- Errors are additive and uncorrelated
- Level
- Intermediate
-
Model 47Validated
Vector Autoregression (VAR)
Models several series jointly, with lag selection, stability, Granger tests, impulse responses, variance decomposition and forecasts.
- Category
- Time Series & Financial Forecasting
- Needs
- Two to four series
- Data
- Time series
- Assumes
- The system is stable
- Level
- Advanced
-
Model 48Validated
Granger Causality
Tests whether lagged values of one series improve the prediction of another, in both directions.
- Category
- Time Series & Financial Forecasting
- Needs
- Two series; Maximum lag
- Data
- Time series
- Assumes
- Both series are stationary
- Level
- Intermediate
-
Model 49Validated
Johansen Cointegration
Tests the number of long-run relationships among several non-stationary series.
- Category
- Time Series & Financial Forecasting
- Needs
- Two to six series
- Data
- Time series
- Assumes
- Series are integrated of order one
- Level
- Advanced
-
Model 50Validated
VECM
Models cointegrated series, reporting long-run relations, adjustment speeds and short-run dynamics.
- Category
- Time Series & Financial Forecasting
- Needs
- Two to four cointegrated series
- Data
- Time series
- Assumes
- Series are I(1) and cointegrated at the chosen rank
- Level
- Advanced
-
Model 51Validated
ARDL & Bounds Test
Estimates an ARDL in error-correction form, with the Pesaran-Shin-Smith bounds test, long-run coefficients and short-run dynamics.
- Category
- Time Series & Financial Forecasting
- Needs
- Dependent series; Independent series; Lags and case
- Data
- Time series
- Assumes
- No series is I(2)
- Level
- Advanced
-
Model 52Validated
ARCH-LM Test
Tests for autoregressive conditional heteroskedasticity in a series or its residuals.
- Category
- Time Series & Financial Forecasting
- Needs
- One series or residual series
- Data
- Time series
- Assumes
- The mean is correctly modelled
- Level
- Intermediate
-
Model 53Validated
GARCH(1,1)
Estimates time-varying volatility with enforced parameter constraints and robust standard errors.
- Category
- Time Series & Financial Forecasting
- Needs
- One return series, 250 to 5,000 observations
- Data
- Time series
- Assumes
- The conditional mean is constant
- Level
- Advanced
Financial & Investment Analysis
-
Model 54Validated
Portfolio Return & Risk
Calculates arithmetic and geometric returns, expected portfolio return, variance, covariance and correlation for chosen weights.
- Category
- Financial & Investment Analysis
- Needs
- Two or more return or price columns
- Data
- Numeric
- Assumes
- Historical means and covariances represent the future
- Level
- Intermediate
-
Model 55Validated
Portfolio Optimization
Finds minimum-variance, global minimum-variance and maximum-Sharpe portfolios under long-only and weight constraints, and verifies convergence.
- Category
- Financial & Investment Analysis
- Needs
- Two or more return or price columns
- Data
- Numeric
- Assumes
- Returns are described by means and covariances
- Level
- Advanced
-
Model 56Validated
CAPM
Estimates beta and alpha by regression and calculates the CAPM expected return.
- Category
- Financial & Investment Analysis
- Needs
- Asset returns; Market returns; Risk-free rate
- Data
- Numeric
- Assumes
- Returns are linearly related to the market
- Level
- Beginner
-
Model 57Validated
Fama-French Three-Factor Model
Estimates exposure to market, size and value factors using factor data that you provide.
- Category
- Financial & Investment Analysis
- Needs
- Asset returns; Mkt-RF, SMB and HML columns that you supply
- Data
- Numeric
- Assumes
- Factor data are aligned and in matching units
- Level
- Advanced
-
Model 58Validated
Sharpe Ratio
Calculates excess return divided by volatility, with annualization and an approximate standard error.
- Category
- Financial & Investment Analysis
- Needs
- Return column
- Data
- Numeric
- Assumes
- Returns are roughly independent for annualization
- Level
- Beginner
-
Model 59Validated
Treynor Ratio
Calculates excess return per unit of market beta.
- Category
- Financial & Investment Analysis
- Needs
- Asset returns; Market returns
- Data
- Numeric
- Assumes
- The benchmark represents market risk
- Level
- Beginner
-
Model 60Validated
Jensen's Alpha
Estimates return above or below the CAPM benchmark, with a t-test and confidence interval.
- Category
- Financial & Investment Analysis
- Needs
- Asset returns; Market returns
- Data
- Numeric
- Assumes
- Well-behaved regression errors
- Level
- Intermediate
-
Model 61Validated
Beta Estimation
Estimates beta against a benchmark across the sample or a rolling window.
- Category
- Financial & Investment Analysis
- Needs
- Asset returns; Market returns
- Data
- Numeric
- Assumes
- A linear relationship with the benchmark
- Level
- Beginner
-
Model 62Validated
Value at Risk (VaR)
Estimates the loss not exceeded at a chosen confidence level, by historical simulation, normal formula and Monte Carlo.
- Category
- Financial & Investment Analysis
- Needs
- Return column
- Data
- Numeric
- Assumes
- The past distribution represents the future
- Level
- Intermediate
-
Model 63Validated
Conditional VaR (CVaR)
Estimates the average loss at or beyond VaR, with the conventions stated.
- Category
- Financial & Investment Analysis
- Needs
- Return column
- Data
- Numeric
- Assumes
- The past distribution represents the future
- Level
- Intermediate
-
Model 64Validated
Monte Carlo Simulation
Simulates correlated asset growth and reports outcome distributions and scenario probabilities.
- Category
- Financial & Investment Analysis
- Needs
- Return, volatility and correlation assumptions
- Data
- Assumptions
- Assumes
- Returns are lognormal with constant parameters
- Level
- Intermediate
-
Model 65Validated
DCF Valuation
Values a business from projected free cash flows, with a terminal value, per-share value and sensitivity table.
- Category
- Financial & Investment Analysis
- Needs
- Free cash flows; Discount rate; Terminal assumption
- Data
- Assumptions
- Assumes
- Cash flows are unlevered
- Level
- Intermediate
-
Model 66Validated
Bond Valuation & Duration
Calculates price or yield to maturity, Macaulay and modified duration, convexity and price sensitivity.
- Category
- Financial & Investment Analysis
- Needs
- Face value; Coupon; Maturity; Yield or price
- Data
- Assumptions
- Assumes
- Coupons are paid in full and on time
- Level
- Intermediate
-
Model 67Validated
Sensitivity & Scenario Analysis
Evaluates your own formula across one-variable, two-variable and base, optimistic and pessimistic scenarios.
- Category
- Financial & Investment Analysis
- Needs
- A formula; Base, low and high values
- Data
- Assumptions
- Assumes
- The formula represents the decision
- Level
- Beginner
Data Analytics & Machine Learning
-
Model 68Validated
Linear Regression Prediction
Evaluates linear regression as a predictor, with predictions, residuals and metrics on held-out data.
- Category
- Data Analytics & Machine Learning
- Needs
- Numeric target; Numeric features
- Data
- Numeric
- Assumes
- Future data resemble the training data
- Level
- Beginner
-
Model 69Validated
Decision Tree
Builds a tree of simple rules, with readable output and feature importance.
- Category
- Data Analytics & Machine Learning
- Needs
- Target; Numeric features
- Data
- Numeric, Categorical
- Assumes
- Future data resemble the training data
- Level
- Intermediate
-
Model 70Validated
Random Forest
Averages many decision trees to improve prediction, with configurable size and feature sampling.
- Category
- Data Analytics & Machine Learning
- Needs
- Target; Numeric features
- Data
- Numeric, Categorical
- Assumes
- Future data resemble the training data
- Level
- Intermediate
-
Model 71Validated
K-Nearest Neighbours
Predicts using the K most similar training rows, with a choice of distance and scaling.
- Category
- Data Analytics & Machine Learning
- Needs
- Target; Numeric features
- Data
- Numeric
- Assumes
- Future data resemble the training data
- Level
- Beginner
-
Model 72Validated
Naive Bayes
Classifies using Bayes' rule with conditionally independent features.
- Category
- Data Analytics & Machine Learning
- Needs
- Categorical target; Numeric features
- Data
- Numeric
- Assumes
- Features are conditionally independent given the class
- Level
- Beginner
-
Model 73Validated
Support Vector Machine
Fits maximum-margin models with linear, RBF or polynomial kernels.
- Category
- Data Analytics & Machine Learning
- Needs
- Target; Numeric features
- Data
- Numeric
- Assumes
- Future data resemble the training data
- Level
- Advanced
-
Model 74Validated
K-Means Clustering
Groups rows around K centroids, with inertia, assignments and an elbow plot.
- Category
- Data Analytics & Machine Learning
- Needs
- Two or more numeric variables; K
- Data
- Numeric
- Assumes
- Clusters are roughly spherical and similar in size
- Level
- Beginner
-
Model 75Validated
Hierarchical Clustering
Merges rows step by step into a tree, with a choice of linkage and distance, and a dendrogram.
- Category
- Data Analytics & Machine Learning
- Needs
- Two or more numeric variables
- Data
- Numeric
- Assumes
- A tree is a sensible summary of the data
- Level
- Intermediate
-
Model 76Validated
Principal Component Analysis
Summarises correlated variables with components, reporting eigenvalues, eigenvectors, explained variance, scores and loadings.
- Category
- Data Analytics & Machine Learning
- Needs
- Two or more numeric variables
- Data
- Numeric
- Assumes
- Linear combinations capture the important structure
- Level
- Intermediate
-
Model 77Validated
Anomaly Detection
Scores how easily each row can be isolated, and flags the highest scores for review.
- Category
- Data Analytics & Machine Learning
- Needs
- Two or more numeric variables
- Data
- Numeric
- Assumes
- Anomalies are few and different
- Level
- Intermediate
-
Model 78Validated
Model Evaluation Toolkit
Evaluates a model with a train/test split and cross-validation, using metrics that match the task.
- Category
- Data Analytics & Machine Learning
- Needs
- Target; Numeric features
- Data
- Numeric, Categorical
- Assumes
- Future data resemble the training data
- Level
- Intermediate
No model matches these filters.