2VhddlmZddlmZddlmZddlmZddlmZedddZ eddd Z ed Gd d ejZ ed GddejZ y))backend)ops) keras_export)squeeze_or_expand_to_same_rank)reduction_metricsz!keras.metrics.pearson_correlationctj|}tj||j}t||\}}|tj||dz }|tj||dz }|tj ||dz }|tj ||dz }tj||z|S)aComputes the Pearson coefficient between labels and predictions. Formula: ```python loss = mean(l2norm(y_true - mean(y_true) * l2norm(y_pred - mean(y_pred))) ``` Args: y_true: Tensor of true targets. y_pred: Tensor of predicted targets. axis: Axis along which to determine similarity. Defaults to `-1`. Returns: Pearson Correlation Coefficient tensor. Example: >>> y_true = [[0, 1, 0.5], [1, 1, 0.2]] >>> y_pred = [[0.1, 0.9, 0.5], [1, 0.9, 0.2]] >>> loss = keras.losses.concordance_correlation( ... y_true, y_pred, axis=-1 ... ).numpy() [1. 0.99339927] dtypeTaxiskeepdimsr )rconvert_to_tensorr rmeanstd)y_truey_predr y_true_norm y_pred_norms U/home/dcms/DCMS/lib/python3.12/site-packages/keras/src/metrics/correlation_metrics.pypearson_correlationrs6 " "6 *F  " "6 >F3FFCNFF388FEEK388FEEK $ NNK $ NNK 88K+-D 99z%keras.metrics.concordance_correlationctj|}tj||j}t||\}}tj||d}tj||d}tj ||z |d}tj ||z |d}||z ||z z}||ztj ||z z}tjd|z|tjzz |S)aComputes the Concordance coefficient between labels and predictions. Formula: ```python loss = mean( 2 * (y_true - mean(y_true) * (y_pred - mean(y_pred)) / ( var(y_true) + var(y_pred) + square(mean(y_true) - mean(y_pred)) ) ) ``` Args: y_true: Tensor of true targets. y_pred: Tensor of predicted targets. axis: Axis along which to determine similarity. Defaults to `-1`. Returns: Concordance Correlation Coefficient tensor. Example: >>> y_true = [[0, 1, 0.5], [1, 1, 0.2]] >>> y_pred = [[0.1, 0.9, 0.5], [1, 0.9, 0.2]] >>> loss = keras.losses.concordance_correlation( ... y_true, y_pred, axis=-1 ... ).numpy() [0.97560976 0.98765432] r Tr r) rrr rrvarsquarerepsilon) rrr y_true_mean y_pred_mean y_true_var y_pred_varcovarnorms rconcordance_correlationr$0s> " "6 *F  " "6 >F3FFCNFF((6t>> pcc = keras.metrics.PearsonCorrelation(axis=-1) >>> y_true = [[0, 1, 0.5], [1, 1, 0.2]] >>> y_pred = [[0.1, 0.9, 0.5], [1, 0.9, 0.2]] >>> pcc.update_state(y_true, y_pred) >>> pcc.result() 0.9966996338993913 Usage with `compile()` API: ```python model.compile(optimizer='sgd', loss='mean_squared_error', metrics=[keras.metrics.PearsonCorrelation()]) ``` cPt|t|||||_d|_yN)fnnamer r up)super__init__rr _directionselfr*r r __class__s rr-zPearsonCorrelation.__init__s3 "   rcJ|j|j|jdSNr*r r r4r0s r get_configzPearsonCorrelation.get_config!IIZZII  r)rN__name__ __module__ __qualname____doc__r-r6 __classcell__r1s@rr&r&_s!J#    rr&z$keras.metrics.ConcordanceCorrelationc0eZdZdZ dfd ZdZxZS)ConcordanceCorrelationaCalculates the Concordance Correlation Coefficient (CCC). CCC evaluates the agreement between true values (`y_true`) and predicted values (`y_pred`) by considering both precision and accuracy. The coefficient ranges from -1 to 1, where a value of 1 indicates perfect agreement. This metric is useful in regression tasks where it is important to assess how well the predictions match the true values, taking into account both their correlation and proximity to the 45-degree line of perfect concordance. Args: name: (Optional) string name of the metric instance. dtype: (Optional) data type of the metric result. axis: (Optional) integer or tuple of integers of the axis/axes along which to compute the metric. Defaults to `-1`. Example: >>> ccc = keras.metrics.ConcordanceCorrelation(axis=-1) >>> y_true = [[0, 1, 0.5], [1, 1, 0.2]] >>> y_pred = [[0.1, 0.9, 0.5], [1, 0.9, 0.2]] >>> ccc.update_state(y_true, y_pred) >>> ccc.result() 0.9816320385426076 Usage with `compile()` API: ```python model.compile(optimizer='sgd', loss='mean_squared_error', metrics=[keras.metrics.ConcordanceCorrelation()]) ``` cPt|t|||||_d|_yr()r,r-r$r r.r/s rr-zConcordanceCorrelation.__init__s3 &   rcJ|j|j|jdSr3r4r5s rr6z!ConcordanceCorrelation.get_configr7r)r$Nr8r9r?s@rrArAs"L'    rrAN)r8) keras.srcrrkeras.src.api_exportrkeras.src.losses.lossrkeras.src.metricsrrr$MeanMetricWrapperr&rArrrJs-@/12$:3$:N56+G7+G\019 *<<9 29 x45: .@@: 6: r