added loglikelihood_lognormal, precompilation true

src/PiecewiseInference.jl

@@ -1,4 +1,4 @@
-__precompile__(false)
+__precompile__()
 """
 $(DocStringExtensions.README)
 """
@@ -52,5 +52,6 @@ module PiecewiseInference
     export piecewise_loss
     export inference, piecewise_ML_indep_TS, iterative_inference, get_ranges
     export plot_convergence
-    export FIM_strouwen, FIM_yazdani, loglikelihood, estimate_σ, RSS, R2, pretty_print, loss_param_prior_from_dict, get_evidence
+    export FIM_strouwen, FIM_yazdani, loglikelihood, estimate_σ, RSS, R2, pretty_print, loss_param_prior_from_dict, get_evidence,
+            loglikelihood_lognormal
 end # module

src/inference.jl

@@ -221,7 +221,7 @@ function inference(infprob;
         u0s_init = _init_u0s(data, ranges)
     end
     # build θ, which is the parameter vector containing u0s, in the parameter space
-    θ = _build_θ(get_p(infprob), u0s_init, infprob)
+    θ = _build_θ(p0, u0s_init, infprob)
 
     # piecewise loss
     function _loss(θ, idx_rngs)

src/statistics.jl

@@ -26,6 +26,43 @@ function RSS(res::InferenceResult, data_set::Array, noisedistrib::T) where T <:
     return rss
 end
 
+"""
+$(SIGNATURES)
+
+calculates the log-likelihood of `data` under the assumption that it follows a Log-Normal distribution. 
+The Log-Normal parameters are derived from `preds`, which represents predicted values, and `σ`, which is the standard deviation.
+
+# Arguments
+- `preds::Array`: A 2D array of predicted values.
+- `data::Array`: A 2D array of observed data, corresponding to the predictions.
+- `σ::Number`: The standard deviation of the Log-Normal distribution, which variance-covariance matrix is assumed as `σ^2 I`
+
+# Returns
+- `Number`: The log-likelihood of the data.
+
+# Notes
+- The function only computes log-likelihood for elements where both predicted and observed data are positive.
+
+# Example
+```julia
+preds = [1.0 2.0; 3.0 4.0]
+data = [1.1 1.9; 2.9 4.1]
+σ = [0.5, 0.5]
+log_likelihood = loglikelihood_lognormal(preds, data, σ)
+"""
+function loglikelihood_lognormal(preds::Array, data::Array, σ::Number)
+    l = 0.
+    for i in 1:size(preds,2)
+        for j in 1:size(preds,1)
+            if preds[j,i] > 0. && data[j,i] > 0.
+                l += logpdf(LogNormal(log.(preds[j,i]), σ), data[j,i])
+            end
+        end
+    end
+    return l
+    # NOTE: pdf(MvNormal(zeros(2), σ^2 * diagm(ones(2))), [0.,0.]) ≈ pdf(Normal(0.,σ),0.)^2
+end
+
 # """
 # $(SIGNATURES)