Search Documents

By S. Shi, X. Xia, F. Xuesai, Z. Shen, D. Han

The hydrodynamic characteristics of a mechanical flotation cell’s transportation zone have significant impact on its metallurgical performance, especially on coarse particles recovery efficiency. However, how to define and adjust the transportation zone was seldom investigated by researchers. In this paper a method will be introduced to define the boundary of the transportation zone by PIV measurement. It is found that the transportation zone can be extended upwards by enlarging the entire rotor-stator assembly from the bottom of tank and increasing the jet angle of the impeller. Two control methods have been validated in industry tests on KYF-160m3 and KYF-320m3 air-forced mechanical flotation cells. The result shows coarse particles recovery are both improved distinctly. It is suggested that enlarging the top transport boundary is able to helpful to reduce the floating distance of bubbleparticle aggregates, which contribute to improve coarse particle recovery in large flotation cells. Key Words: Flotation cell; Transportation zone; Coarse particles flotation

Jan 1, 2020

By T. Bhambhani, J. Jankolovits

Sphalerite flotation using copper sulfate has been performed in the industry for decades and the mechanism of Cu activation is well established. Despite its ubiquity, many practical questions surrounding copper sulfate persist. It is not clear how iron selectivity is achieved in zinc circuits given the propensity of iron sulfides to be activated by copper. Also, the large variation in copper sulfate dosages across different plants and ore types cannot be explained. Lastly, Zn flotation with greatly reduced copper sulfate dosages attracts significant interest and has been demonstrated in certain laboratory settings, but no information is available on why those laboratory solutions have not been adopted in plant practice. Flotation studies are presented that demonstrate that Zn flotation with very low copper sulfate dosages is possible when using certain novel collectors that are not practised in the industry. However, these reagent schemes are impractical because of a heretofore unappreciated and critical role that copper sulfate plays in Fe rejection. With both very low and standard copper sulfate dosages, copper sulfate influences both Zn flotation and Fe rejection. The results highlight the multifaceted role of copper sulfate in zinc flotation and the need to holistically consider complex systems’ behaviour to understand flotation reagent mechanisms and optimise plant performance. Keywords: Sphalerite flotation, metal activation, copper sulfate, iron rejection

Jan 1, 2020

By D. W. Dean, J. A. Ryder

"A computer system designed to assist in maintenance scheduling and costing, and in monitoring the performance of equipment, is described. The emphasis is on the information retrieval and implementation aspects of the system.INTRODUCTIONThe cost of the maintenance of equipment can account for a substantial proportion of the total costs of any large industry. These costs are made up of the following elements:(i) capital expenditure,(ii) maintenance labour,(iii) maintenance spares,(iv) maintenance overheads,(v) production losses.Control of these cost elements cannot be achieved without a system which maintains and retrieves the necessary infor¬mation efficiently.The system should provide engineering management with the information necessary to:(i) establish the cost of maintaining items of equipment,(ii)establish the economic life of equipment,(iii) determine breakdown patterns for the establishment of optimum inspection, service and overhaul frequencies,(iv) establish work schedules for those responsible for the maintenance of equipment,(v) establish the optimum maintenance labour complement,(vi) establish the optimum stockholding of maintenance stores items,(vii) improve the performance of equipment,(viii) justify the selection of a particular type and make of equipment,(ix) establish the reasons for any abnormal usage of materials and labour, and to(x)reduce the dependence on the human element for the control of equipment maintenance activities.Many industries have comprehensive and numerous equipment maintenance records but they experience great difficulty in retrieving the required information in the time available and with the labour available. Many industries do not bother to maintain meaningful records, because of this difficulty."

Jan 1, 2014

By E. J. Y., G. J. McLachlan, E. Amini, Koha

Inferring individual component flotation rates from recovery-time data of ore complexes is important in optimising and designing minerals processing plants. However, existing two-component first-order flotation models are computationally intensive to solve using conventional non-linear least squares regression. In this paper, a Convolutional Neural Network (CNN) is used to approximate the inverse function which predicts the two-component first-order model parameters using recovery-time data. This paper compares the accuracy and computation speed of the proposed CNN against traditional non-linear least squares regression on 1 million generated data based on a known two-component firstorder kinetic model. It is shown that using a CNN is possible to approximate the inverse function for the non-invertible two-component model with recovery-time as an input matrix for convolution operations. The CNN method is 450 times faster to compute than the Trust-Region (TR) non-linear least squares regression algorithm. However, the CNN model is less accurate compared to the TR algorithm overall. Keywords: Flotation kinetics, convolutional neural network, inverse problem, modelling and simulations

Jan 1, 2020

By X. Zhang, Z. T. Yuan, C. Zhang, H. M. Zheng, X. M. Wang, L. X. Li, N. Li

In metal mines, comminution is commonly used to liberate minerals for subsequent mineral concentration processes, such as flotation, magnetic separation and gravity separation. However, ores are aggregates of valuable minerals and their disseminated gangue minerals, and these minerals exhibit diverse comminution behaviours. In carbonate-bearing iron ores, hematite, siderite, and quartz are the predominant minerals, of which hematite is the objective recoverable mineral. In this regard, the present study aimed to identify the imperfections in siderite minerals by X-ray computed tomography (XCT), and thus quantitatively characterise their size and shape in terms of Feret diameter and aspect ratio, as well as other parameters. XCT images indicated the protogenetic ubiquity of pores <5 μm in diameter and cracks of 7-8 μm inside siderite particles. In the ground product of artificially mixed siderite and hematite, siderite particles were smaller than hematite particles. Siderite particles had a larger surface area and greater surface roughness than hematite particles of the same volume, which was attributable to the intrinsic holes and cracks in siderite particles. This study provides an insight into the relationship between the inherent imperfections of minerals and their comminution properties. Keywords: Siderite; X-ray computed tomography; particle imperfections; grinding behaviour; sliming

Jan 1, 2020

By L. A. Maseko

"As part of the mining engineering curriculum, students are required to be placed in industry for vacation work or work-integrated learning (WIL). In this paper we review work-based placement as done at the four South African mining schools: the University of Johannesburg, the University of Pretoria, the University of the Witwatersrand, and the University of South Africa.Key aspects of past and current WIL (pre- and post-graduation) practices that may need attention are highlighted for consideration by all universities that provide professional qualifications. Work-directed theoretical learning, problem-based learning, project-based learning, workplace learning, and simulated work experiences are discussed as legitimate approaches to WIL.Regardless of the choice each institution makes, commitments of universities, preparedness of university teachers, and the desire for equity and excellence will play an important role in the provision of WIL.DefinitionsAccredited Qualification: a qualification awarded on successful completion of an accredited programme.Advisory Boards (or Committees): are essential infrastructure for a WIL curriculum that is built on the integration of the workplace and the university. This requires the collaboration of experts from both environments. Workplace advisors are critical to the success of the academy as they embody expert knowledge of the ‘real world’.Assessment: the process of determining the capability or competence of an individual against standards by evaluating performance.Category: a mode of registration defined in or under the ECSA Act which has a distinctive purpose, characteristic competencies, defined principal routes to registration, and educational requirements.Experiential learning: (in South Africa sometimes used synonymously with work placement or cooperative education) is a term used with a great variety of meanings in the international literature. More broadly, it may refer to learning that has meaningful learner involvement. It is the process of generating meaning from direct experience. In the international literature, experiential learning is learning through reflection on doing, which is often contrasted with rote or didactic learning."

Dec 1, 2018

By F. P. Janse van Rensburg, P. J. Le Roux

"There are numerous factors that affect open stope stability and often result in falls of ground. These falls of ground can be attributed to factors such as beam failure due to a larger than normal roof area (hydraulic radius too large), adverse ground conditions, seismicity, the stress-strain environment, absence of support, and poor drill-and-blast practices. The effect of time on the stability of open stopes is sometimes underestimated and is relatively unknown, especially on Target Mine. Actual data was collected from open stopes at Target and analysed to show the effect of time on open stope stability. The benefits of this analysis will include improved understanding of time-dependent failure, which can assist in reducing dilution and the risk of sterilization of future mining blocks. IntroductionTarget Mine is situated adjacent to the town of Allanridge, some 20 km from Welkom in South Africa’s Free State Province, and is the most northerly mine in the Welkom Goldfields area (Figure 1). The mine consists of a single surface shaft system with a sub-shaft (Target 1C shaft) and a decline. Ownership was attained by Harmony Gold Mining Company Limited in May 2004 (Harmony Gold Mining Company Limited, 2010). Before discussing the selection of open stopes for the back-analyses, a brief explanation of open stoping as practiced at Target Mine will be given. The orebody is some 180 m in thickness and 270 m wide and comprises multiple reefs overlying one another. The 180 m thick reef package is termed the Eldorado Reefs, as shown in Figure 2. The Eldorado Reefs suboutcrop against the Dreyerskuil Reefs, as shown in Figure 3. The dip of the reef varies from as low as 10° in the west to 75° in the east.Compared with most Australian and Canadian open stoping mining operations, Target Mine is unique. In most Australian and Canadian mining operations the hangingwall and footwall of the open stopes comprise waste rock, and the orebody dips relatively steeply. Due to the depth of Target Mine, some 2300 m to 2500 m below surface, a destress slot (Figure 4) is mined to create an artificial, shallow mining environment whereby the field stress is managed and kept at values of around 60 MPa. The destress slot comprises a narrow tabular stope with an average width of 1.5 m, and is mined on the Dreyerskuil Reefs.At Target Mine the hangingwall, sidewalls, and footwall of the open stopes all comprise reef of different grades, except for the EA1, where the EB footwall is waste rock. If a stope, for example, is mined alongside an existing old stope, the western sidewall of this stope will consist of backfill. The general mining direction of open stopes is from the lowest position of the reef on the west, progressing up towards the east as shown in Figure 5."

Apr 1, 2019

By X. Yang, G. D. Zhai

The prop is the most important part to ensure the proper functioning of hydraulic support. When the hydraulic support is impacted by the roof, the prop is prone to extrusion deformation, expansion, and even bursting. In order to study the stress on a doubly-telescopic hydraulic support prop under impact load, an impact simulation was carried out based on the drop weight method. First, the impact model of the hydraulic system of the prop was established in AMESim software and the dynamic response curves of the bottom and middle cylinder of the prop obtained. Then, according to the conservation of energy, the pressure formulaes in the hydraulic cylinder of the prop under impact load were derived and verified by the AMESim simulation results, and the maximum pressure on the inner surface of the hydraulic cylinder of the prop obtained. Lastly, the transient dynamic simulation of a hydraulic support prop was carried out in ANSYS Workbench software, and cloud diagrams of stress and deformation of the prop obtained using Workbench simulation. The AMESim simulation process describes the change in fluid pressure in the prop, and the pressure formulae can be used to estimate the internal pressure of the prop under impact load. The finite element analysis results show that the stress of the middle hydraulic cylinder is much greater than that on the bottom hydraulic cylinder under impact load, which can provide a reference for prop design.

Jul 2, 2022

By A. J. S. Spearing, M. Sharifzadeh, K. V. Jessu

Empirical pillar design methods are commonly used in the mining industry. The parameters within which these methods are valid are frequently unknown to the user or ignored. In addition, empirical design may not consider essential parameters such as blasting effects, orebody dip and the presence of geological structures, which all adversely affect the stability of the pillars. This can result in potentially serious pillar design strength over-estimates. Although the commonly based tributary area method is generally conservative, as the spans are seldom that large relative to the depth, failing to consider other relevant parameters can result in errors. Problems associated with an under-designed pillar can range from a local pillar collapse to a catastrophic chain reaction collapse (or run). Over-designed pillars are generally safe but reduce the extraction of the orebody, thus adversely affecting the profitability of the mining operation. We used laboratory tests and numerical modelling to understand the effects of pillar orientation, blasting and the presence of discontinuities on pillar strength. Reduction factors were developed with these models to be implemented in conjunction with the existing empirical pillar design methods. For any pillar or mine design, once it is implemented, the actual performance of the system must be checked regularly by observation and monitoring and adjusted if needed. The pillar design approach outlined in this paper can better optimize the pillar mining method by considering other generally ignored but important parameters, thus improving safety, productivity, and economic aspects.

Oct 1, 2022

By L. Zhou, J. Qiu, M. Tang, J. Wang, G. Dai

Coal spontaneous combustion in longwall gobs is a major concern in coal mines. In comparison to other hazards, this phenomenon is difficult to detect. Fifteen factors influencing coal spontaneous combustion in longwall gobs were identified from the literature. The logical relationships between these factors were evaluated and quantified by decision-making trial and evaluation laboratory (DEMATEL) and interpretive structural modelling (ISM) methods. A four-level hierarchical structure model was established. The research shows that combining DEMATEL and ISM can reveal the key factors that influence coal spontaneous combustion in longwall gobs and establishthe interactions and relationships among these factors. The results can provide the basis for preventing and controlling coal spontaneous combustion in longwall gobs.

Jun 10, 2024

"Chairman: Professor J. de V. Lambrechts Rapporteur: Mr R. HempPapers:Computer-aided design of a flash smelting installation by M. A. T. Cocquerel, R. T. Gates and L. A. GreenPlanning of mine ventilation by E. U. Reuther and A. DohrnenCoal mine ventilation systems in the U.S.A. by R. Stefanko and R. V. Ramani Heat flow models in ventilation planning by A. J. Dickson and A. M. StarfieldThe first of the papers in the session dealt with a flash smelter installation, while the remaining three papers dealt with mine ventilation. In view of this the Chairman proposed that the paper on flash smelting be introduced and discussed first, after which the remaining three papers would be introduced followed by discussion on all three.Dr Cocquerel, who presented the first paper, used some typical flow diagrams to illustrate the process and stated that the program considered only steady-state operations and did not take into account the dynamical changes. Subjective judgment was necessary for the preparation of input data and some degree of user appreciation was required.Dr Cocquerel said that no work bad yet been done on an optimization program, but that this was being considered. Such a program would require a large amount of physical and chemical data which were not yet available. The variables to be optimized were the energy and material consumptions and the loss of metal in the slag.The Chairman opened the discussion by asking Dr Cocquerel if he could give a brief description of the flash smelting process for the benefit of those not familiar with the subject. Dr Coquerel replied that in the Outokumpu process con¬centrates were injected directly into the furnace, and that the process was somewhat analogous to flash distillation or evaporation processes, but was perhaps closer to the burning of pulverized fuel.The process had two distinct advantages. First, the heat evolved in the smelting reaction was used to heat the furnace, which resulted in less fuel consumption, and, second, the effluent gases were cheaper to treat than the effluent gases from other pyrometallurgical processes. This was important particularly in view of the current interest in pollution problems.Mr P. Marais thought that the authors had done themselves an injustice by not including details of the very complex processes taking place in the furnace."

Jan 1, 2014

By D. F. Malan, T. E. Oates

A recent experimental pillar extraction project at a UG2 bord-and-pillar mine presented a unique opportunity to compile a new pillar database. Currently, the South African hard rock bordand- pillar mines are designed using the Hedley and Grant formula with a modified K-value. This empirically derived formula was developed for uranium mines in the Elliot Lake district of Canada. The use of this formula for the design of pillars in South Africa is questionable. Very few pillar failures have nevertheless been observed and its current calibrations for the various reef types are possibly too conservative. A new UG2 pillar database of 66 pillars, of which seven are classified as failed, was compiled by the authors. This enabled a revised ‘first-order’ calibration of the K-value for the Hedley and Grant formula. The new estimated value for the UG2 is K = 75 MPa. This gives a pillar strength that is more conservative than the PlatMine formula. This work should nevertheless be considered as only a preliminary calibration as the database was small. Further work is also required to determine whether the exponents in the formula for the width and height parameters are appropriate for UG2 pillars.

May 3, 2023

By G. Segaetsho

"Conventional blasting rules of thumb used to develop blast designs are based on parametric ratios, with insufficient consideration of the rock mass factors that influence the achievability of final wall designs. At a hard-rock mine in Botswana, catchment berm designs were frequently unachievable, which exposed the mining operation to safety hazards such as local wallrock failure from damaged crests, frozen toes, and rolling rockfalls from higher mining levels. Increased standoff distances from the highwall reduced the manoeuvring area on the pit floor and the extraction factor that was safely achievable. Application of rock mass classification and the Blastability Index (BI) as a means to improve wall control was investigated. Zones were defined according to rock types forming the western highwall rock mass, and distinguishing rock mass classification factors were used to establish the suitable wall control designs. The approach combined rock mass classification methodologies such as the Geological Strength Index and the Rock Mass Rating, and related them to the BI and discontinuities of the rock mass to produce a design input tool that can be used to develop objective wall control designs. The designs driven by the tool inherently take into account the rock mass characteristic factors at the centre of the rock mass classification methods mentioned above, and significantly reduce the dependence on rules of thumb. It was found that this approach produces designs with explosive energy measures that are consistent with the required rock-breaking effort and the behaviour of discontinuities, while preserving perimeter walls.IntroductionBlasting is the first step in the comminution process. Various coordinated inputs contribute to its success, and in turn to the efficient execution of processes that lie downstream from the blasting step. Primary to meeting production and operational objectives of a mine, as well as the protection of human life and the mineral resource, is the achievement of highwall design specifications and competence. To this end, this study investigated the use of existing tools and knowledge for the achievement of highwall control goals at a hard-rock mine in Botswana."

Jan 1, 2018

By T. Nghipulile, M. Moys, M. Bwalya, H. Simonsen

The quest for efficiency in comminution is an ongoing concern as comminution usually constitutes a major cost component in the metal production industry. Such improvements can be made by circuit optimization or development of more efficient equipment. A novel crusher, known as the rotary offset crusher (ROC), promises to deliver in this space. The ROC was invented in 2002 by Michael Hunt, Henry Simonsen, and Ian Sinclair, but failed to garner enough support to progress to production. The original design concept was recently rekindled, and a laboratory version of the crusher has been built and commissioned at the University of the Witwatersrand. The crusher is simple in design, with two cylindrical discs that are parallel to each other, and, as the name implies, there is an offset between the vertical axes of the discs. The top disc has a conical section on its lower face, and this creates a crushing zone between the opposite faces of the two spinning discs. Centrifugal motion transports particles through the crushing zone. Batch experiments have been conducted with quartz at various crusher settings (discs offset, rotational speed, and vertical exit gap) for various feed size distributions. The indications so far suggest that the disc speed is a key factor affecting the performance. Size reduction ratios as high as 11 were recorded from experiments with quartz at a speed of 830 r/min.

Jan 1, 2021

By S. J. Webber

This paper reports on the use of the MINSIM-D stress-analysis program for the modeling of two large seismic events that occurred on the same fault eleven years apart. It was found that the first event (magnitude 5,2) could not be ascribed to purely mining-induced stresses, indicating that the largest seismic events probably result from the interaction of mining-induced stresses with large scale tectonic stresses. However, the stress changes induced during eleven years of mining were found to be capable of generating sufficient excess shear stress (ESS) to be responsible for the second event (magnitude 4,3), The upper limit for the magnitude of purely mining-induced events was found to be approximately 4,3.

Jan 1, 1990

By S. Annavarapu

"Fragment size distributions play an important part in the design and planning of block cave mining operations. Though several methods have been proposed for estimating fragment size distributions in advance of mining, the calibration of these estimating procedures to field measurements has been a challenge. The Block Size Estimator (BSE) program was developed for providing an assessment of primary fragmentation expected in a block cave mine using drill-core piece length data and joint characteristics from the exploration and geotechnical evaluation programmes. In order to validate the results from the BSE program, the drill-core piece length data and the joint characteristics in different rock types from the DOZ block cave mine in Indonesia were used to help generate fragment size distributions in different areas of the mine. The predicted fragment size distributions were compared with the fragmentation observed at the drawpoints during operation of the mine.This paper presents the details of the validation of the fragment size estimates by the BSE program using fragment size distribution data from the drawpoints at the extraction level at the DOZ block cave mine, and the challenges encountered in developing reasonable correlations between the estimated and measured fragment size distributions. IntroductionThe assessment of fragmentation is an important aspect of the design and planning of the block cave mining method, and forms an integral part of the design of the excavations at the extraction level and the selection of material handling systems for transporting the ore to the processing stations. Secondary blasting requirements can also be estimated based on the fragment size distributions developed for the block cave (Laubscher, 1994).Several methods of estimating fragment size distributions in block cave mines have been developed based on joint set parameters estimated from structural data from oriented core drilling campaigns and mapping of available excavations and outcrops. The BCF program (Esterhuizen, 1999) estimates the fragmentation using a combination of empirical, analytical, and rational methods to model the behaviour of materials during the primary and secondary fragmentation processes. The program JKFrag uses advanced tessellation of joint traces to create rock blocks and then generate fragment size distributions (Brown, 2002). Nickson, Coulson and Hussey (2000) presented the details of processing of the geotechnical data from the Mont Porphyre project, including an assessment of block sizes from the data collected on the core logging sheet. The distance between natural fractures was used to create distributions of block lengths, and a distribution of potential block sizes was estimated assuming that the blocks had equal dimensions in all directions. The results agreed closely with the results of the BCF program. Hadjigeorgiou, Grennon, and Nickson (2002) demonstrated the use of oriented borehole data to provide characteristic block size distributions for a mining project using a software package called Stereoblock, which generates a three-dimensional joint network for a given volume and calculates the volumes of blocks created by the intersections of these joints, simulated as circular planes. The distributions of joint orientation and joint spacing generated from statistical analysis of the oriented core data are used for the creation of the Stereoblock model. The diameter of the circular planes representing the joints is determined based on the mean and standard deviation of the joint trace length obtained from scan-line mapping."

May 1, 2019

By А. А. Vasiliev, Y. G. Seredkin, А. V. Aksenov, А. Y. Senchenko

Refractory gold-containing concentrates usually contain gold enclosed in sulphides as well as in rock-forming minerals. Another difficulty is that gold often occurs as very fine inclusions of several micrometres, and even smaller than 1 micrometre, or can be part of the mineral composition at the structure level. Processing of such concentrates is a challenge. Ultrafine grinding and further oxidation are applied for gold liberation. The paper deals with the results from the study of gold deportment using Qemscan, and atmospheric oxidation testwork involving Albion Process™ on a gold-containing concentrate with a high content of refractory gold in sulphides and gangue minerals. Albion process testing was performed at Institute TOMS with the participation of Glencore Technology Pty Ltd (Australia). Additionally, conventional cyanidation testwork was carried out for the concentrate processing. As an alternative technology of gold liberation, oxidising roasting with arsenic stabilisation was used. The results from the study make it possible to assess the efficiency of the above-mentioned methods of gold liberation, and applicability of Albion technology to recover gold from the refractory concentrate. Key words: Gold, refractory concentrate, Albion Process, atmospheric oxidation, oxidising roasting, leaching

Jan 1, 2020

By M. J. Nicol, G. Cromberge, C. A. Flemming

A relatively simple rate equation for the kinetics of the absorption of gold from cyanide solutions onto activated carbon was lI5ed as the basis in the development of a model for multistage carbon-in-pulp and carbon-in-Ieach absorption circuits. The model was used in the prediction of both the steady-state and the transient behaviour of the absorption process, and the results were found to be in reasonable agreement with those observed in an extended pilot-plant campaign. A number of practical implications of the model are discussed, and an assessment is made of the economic consequences of the various process options available for the design of a carbon-in-pulp plant.

Jan 1, 1983

By F. J. N. Stassen

SYNOPSIS The effect of conditioning energy on the flotation of gold, U308, and pyrite was investigated in the range 0,1 to 100 kWh per tonne of dry ore for various combinations of conditioning time and impeller speed in a cylindrical con¬ditioning tank. It was found that, when the conditioning energy was increased to between 5 and 10 kWh per tonne of dry ore, the total recovery and flotation rate of the valuable minerals (expressed as Klimpel parameters) increas¬ed substantially. The Klimpel parameters are dependent on conditioning energy, but are independent of condition¬ing time or impeller speed (at constant conditioning energy). The Klimpel parameters of the gangue are indepen¬dent of conditioning energy.

Jan 1, 1991

By A. Drumond, J. F. C. L. Costa, M. G. Lemos, A. L. Rodrigues, F. G. Niquini

The past decade has seen a renewed possibility of using machine learning algorithms to solve a large collection of problems in several fields. Data acquisition for mining operations has increased with the growth in sensor-based technologies, and therefore the amount of information available for mining applications has dramatically increased. Ore sorting equipment is available for separating ore from waste based on differences in physical properties detected by a real-time analyser. The separation efficiency depends on the contrast in these properties. In this study we investigate the application of machine learning models trained using data from the output of a dual-energy X-ray ore sorting apparatus at a gold mine. The particles were first hand-sorted into ore and gangue classes based on their mineralogical composition. Classification models were then used to help decide the balance between the number of true and false positives for ore in the concentrate, with a view to economic parameters, using their receiver operator characteristic (ROC) curves. The results showed AUC (area under the ROC curve) scores of up to 0.85 for the classification models and a maximum reward condition Fpr/Tpr around 0.5/0.9 for a simplified economic model.

Aug 8, 2024