Gold deposits in the Jiaodong Peninsula represent a primary gold resource in China and mostly exhibit similar ore-forming features related to sericite-quartz-pyrite alteration and other controls from (micro-)structural deformation. This study investigates the pyrite textures and trace elements in the Sizhuang gold deposit (>100 t Au) to document the key factors impacting on the genesis of the Jiaodong-type orogenic deposits. Three main types of pyrite are identified: (1) the first generation of pyrite (Py1) occurs as disseminated euhedral to subhedral grains in K-feldspar-albite-rutile-hematite and sericite alteration (stage 1), (2) Py2 as aggregates in quartz-sericite-pyrite altered rocks or quartz-pyrite veins (stage 2) can be subdivided into Py2a as irregular cores, Py2b as a zoned overgrowth on Py2a, and Py2c as overgrowth on early pyrite, and (3) Py3 as fine-grained crystals in siderite-polymetallic veins (stage 3). Primary gold at the Sizhuang deposit is coevally or slightly later deposited with Py2b, Py2c, and Py3. Laser ablation–inductively coupled plasma mass spectrometry (LA–ICP–MS) analyses show that the highest Co and Ni contents in Py1 and high but variable Co in Py2b favors the involvement of deep high-temperature magmatic waters at stage 1 and middle stage 2. The elevated As contents from Py2a to Py2c and depletion of trace elements (e.g., Co, Ni, As and Te) and high Au/Co, Cu/Ni, and As/Ni values in Py2a and Py3, combined with published H-O isotope data, imply a meteoric water ingress during stage 2–3. Thus, the fluid evolution at Sizhuang is a consequence of pulsed deep magmatic fluid release plus progressive meteoric fluid ingress. The rhythmic Co–As–Ni–Au bands of Py2b additionally suggest episodic changes in the composition of ore-forming fluids. Moreover, the sharp textural features (e.g., pyrite overgrowth on previously cataclastic crystals) of Py2 and As-Cu-rich and Co-poor bands in zoned Py2b probably also reflect rapid metal deposition and self-organization and subsequent mineral crystal growth due to the pressure release during phase separation in the Sizhuang deposit. Considering the significantly concentrated gold (>1300 t) in the regional Jiaojia fault zone and Au-bearing mineral formation related to phase separation (boiling) in the Sizhuang deposit, gold mineralization in the Sizhuang deposit was interpreted to be controlled by the pressure-driver owing to the seismic activities in the Jiaojia fault system.

The weak classifier ensemble algorithms based on the decision tree model, mainly include bagging (e.g., fandom forest-RF) and boosting (e.g., gradient boosting decision tree, eXtreme gradient boosting), the former reduces the variance for the overall generalization error reduction while the latter focuses on reducing the overall bias to that end. Because of its straightforward idea, it is prevalent in MPM (mineral prospectivity mapping). However, an inevitable problem in the application of such methods is the hyperparameters tuning which is a laborious and time-consuming task. The selection of hyperparameters suitable for a specific task is worth investigating. In this paper, a tree Parzen estimator-based GBDT (gradient boosting decision tree) model (TPE-GBDT) was introduced for hyperparameters tuning (e.g., loss criterion, n_estimators, learning_rate, max_features, subsample, max_depth, min_impurity_decrease). Then, the geological data of the gold deposit in the Xiong ‘ershan area was used to create training data for MPM and to compare the TPE-GBDT and random search-GBDT training results. Results showed that the TPE-GBDT model can obtain higher accuracy than random search-GBDT in a shorter time for the same parameter space, which proves that this algorithm is superior to random search in principle and more suitable for complex hyperparametric tuning. Subsequently, the validation measures, five-fold cross-validation, confusion matrix and success rate curves were employed to evaluate the overall performance of the hyperparameter optimization models. The results showed good scores for the predictive models. Finally, according to the maximum Youden index as the threshold to divide metallogenic potential areas and non-prospective areas, the high metallogenic prospect area (accounts for 10.22% of the total study area) derived by the TPE-GBDT model contained > 90% of the known deposits and provided a preferred range for future exploration work.