Application of Class-Based Machine Learning for Potential Hydrocarbon Zones Identification: A Case Study

At times, petrophysicists are expected to evaluate potential of the well in time-constraint situations while maintaining consistency of the parameters and interpretation. Other than that, some challenges may also occur when working with older wells where the dataset are not as complete as current wells and processing parameters are not transferable. In this case study, class-based machine learning (CBML) approach is used to perform petrophysical evaluation to identify potential hydrocarbon zones in the target wells. The objective is to find solution to improve efficiency and consistency in those challenging situations. A class-based machine learning (CBML) workflow uses cross-entropy clustering (CEC)-Gaussian mixture model (GMM)- hidden Markov model (HMM) workflow that identifies locally stationary zones sharing similar statistical properties in logs, and then propagates zonation information from training wells to other wells (Jain, et al., 2019). The workflow is divided into two (2) main steps: training and prediction. Key wells which best represent the formation in the field are used to train the model. This approach automatically generates the number of cluster (class) using unsupervised or supervised depending on the input data. The model from key wells data is then used to reconstruct inputs and outputs along with uncertainty and outlier flags. This allows expert to QC and validate the generated class which is the most crucial part of the workflow. Once the model from the key wells has been built, it is applied to predict the same set of zones in the new wells that require interpretation and predict output curves. The result matched well over the good data interval with the petrophysical interpretation result from conventional approach. While in the bad interval, some discrepancies can be observed. The discrepancy was identified easily from the uncertainty and outlier flags which helps petrophysicists to identify which interval to fix or re-evaluate. Some requirements to condition the input were observed (no missing value over the input and outlier) to get the best result. A number of inputs used in the model need to be consistent over the set of wells used in the training and prediction target. This machine learning workflow speeds-up the petrophysical analysis process, reduce analyst bias and improve consistency result between one well to another within the same field. This machine learning application can also generate auto log QC, zonation class for rock typing also reconstructed logs which enrich the petrophysical interpretation even for wells with limited logs availability. This paper offers practical examples and lessons learned of CBML approach application to perform petrophysical evaluation and identify potential zones while being in time-constrained and limited resource situations.

Analyst and etching protocol effects on the reproducibility of apatite confined fission-track length measurement, and ambient-temperature annealing at decadal timescales

Previous inter-laboratory experiments on confined fission-track length measurements in apatite have consistently reported variation substantially in excess of statistical expectation. There are two primary causes for this variation: (1) differences in laboratory procedures and instrumentation, and (2) personal differences in perception and assessment between analysts. In this study, we narrow these elements down to two categories, etching procedure and analyst bias. We assembled a set of eight samples with induced tracks from four apatite varieties, initially irradiated between 2 and 43 years prior to etching. Two mounts were made containing aliquots of each sample to ensure identical etching conditions for all apatites on a mount. We employed two widely used etching protocols, 5.0 M HNO3 at 20 °C for 20 s and 5.5 M HNO3 at 21 °C for 20 s. Sets of track images were then captured by an automated system and exchanged between two analysts, so that measurements could be carried out on the same tracks and etch figures, in the same image data, allowing us to isolate and examine the effects of analyst bias. An additional 5 s of etching was then used to evaluate etching behavior at track tips. In total, 8391 confined fission-track length measurements were performed; along with 1480 etch figure length measurements. When the analysts evaluated each other's track selections within the same images for suitability for measurement, the average rejection rate was ~14%. For tracks judged as suitable by both analysts, measurements of 2D and 3D length, dip, and c-axis angle were in excellent agreement, with slightly less dispersion when using the 5.5 M etch. Lengths were shorter in the 5.0 M etched mount than the 5.5 M etched one, which we interpret to be caused by more prevalent under-etching in the former, at least for some apatite compositions. After an additional 5 s of etching, 5.0 M tracks saw greater lengthening and more reduction in dispersion than 5.5 M tracks, additional evidence that they were more likely to be under-etched after the initial etching step. Systematic differences between analysts were minimal, with the main exception being likelihood of observing tracks near perpendicular to the crystallographic c axis, which may reflect different use of transmitted vs. reflected light when scanning for tracks. Etch figure measurements were more consistent between analysts for the 5.5 M etch, though one apatite variety showed high dispersion for both. Within a given etching protocol, each sample reflected a decrease of mean track length with time since irradiation, giving evidence of 0.2–0.3 μm of annealing over year to decade timescales.

Does Forecast Bias Affect Financial Analysts’ Market Influence?

Prior studies find that analysts tend to bias their forecasts upward in poor information environments and downward in rich information environments, consistent with attempts to curry favor with management. We find that investors anticipate this behavior by reducing their response to upward forecasts in poor information environments and downward forecasts in rich information environments. Using Hugon and Muslu’s measure of analyst conservatism as an ex ante indicator of individual analysts’ forecast bias tendencies, we show that the stronger return response they find to conservative analysts’ forecast revisions is restricted to poor information environments, where optimistic analyst bias is prevalent. Our results suggest that analysts pay a price in market influence when their forecasts reinforce analysts’ typical forecast bias for the firm’s information environment. Conversely, analysts whose forecasts conflict with the typical bias for the firm are rewarded with larger than average return responses.

Monitors or Predators: The Influence of Institutional Investors on Sell-Side Analysts

ABSTRACT: Regulators and the investment community have been concerned that institutional investors pressure financial analysts through trading commission fees to issue optimistic opinions in support of their stock positions. We use a unique dataset that identifies mutual fund companies' allocation of trading commission fees to individual brokerages and provide direct evidence on this issue. In particular, we show that for stocks in which the fund companies have taken large positions, analysts are more optimistic in their stock recommendations when their brokerages receive trading commission fees from these fund companies. The relationship is stronger when the commission fee pressure is greater. The market reacts less favorably to the “Strong Buy” recommendations of analysts facing greater commission fee pressure. The funds also respond negatively to such recommendations in making portfolio adjustments. These results point to a source of analyst bias that has been little explored in the literature. Data Availability: The data are publically available from the sources identified in the paper.