37 results about "Selection bias" patented technology

A preferred embodiment of the present invention comprises a method of creating and managing an index fund based on an index of funds of hedge funds, comprising steps of (a) identifying potential funds for an index of funds of hedge funds; (b) identifying which of the potential funds meet specific criteria selected so as to minimize biases comprising selection bias and survivorship bias; (c) creating an index by indexing funds identified as meeting the criteria; (d) acquiring a portfolio of funds in the index; and (e) allocating capital to funds in the portfolio.The invention also comprises software to perform the steps of the method described above. In particular, the invention comprises software wherein data regarding potential funds is stored in a database according to object role modeling. A preferred database structure is described.

A method for managing sample selection bias is disclosed. Embodiments of the method can include automatically determining an unbiased estimate of a distribution from occurrence data via a special purpose processor. The occurrence data that is utilized in determining the estimate may have an occurrence data sample selection bias that is substantially equivalent to a background data sample selection bias associated with background data. Additionally, the occurrence data may be related to the background data, and the background data may be chosen with the background data sample selection bias. Furthermore, the occurrence data may represent a physically-measurable variable associated with one or more physical and tangible objects or substances.

New computational technologies generating systematic investment portfolios by coordinating forecasting algorithms contributed by researchers are provided. Work on challenges is efficiently facilitated by the algorithmic developer's sandbox (“ADS”). Second, the algorithm selection system performs a batch of tests that selects the best developed algorithms, updates the list of open challenges and translates those scientific forecasts into financial predictions. The algorithm controls for the probability of backtest overfitting and selection bias, thus providing for a practical solution to a major flaw in computational research involving multiple testing. Third, the incubation system verifies the reliability of those selected algorithms. Fourth, the portfolio management system uses the selected algorithms to execute investment recommendations. A dynamically optimal portfolio trajectory is determined by a quantum computing solution to combinatorial optimization representation of the capital allocation problem. Fifth, the crowdsourcing of algorithmic investments controls the workflow and interfaces between all of the hereinabove introduced components.

A preferred embodiment of the present invention comprises a method of creating and managing an index fund based on an index of funds of hedge funds, comprising steps of (a) identifying potential funds for an index of funds of hedge funds; (b) identifying which of the potential funds meet specific criteria selected so as to minimize biases comprising selection bias and survivorship bias; (c) creating an index by indexing funds identified as meeting the criteria; (d) acquiring a portfolio of funds in the index; and (e) allocating capital to funds in the portfolio. The invention also comprises software to perform the steps of the method described above. In particular, the invention comprises software wherein data regarding potential funds is stored in a database according to object role modeling. A preferred database structure is described.

The invention aims to provide a method and a device for determining selected texts on touch terminals. A selection device obtains text selection operation executed on a touch terminal, and the text selection operation corresponds to at least one candidate text region which comprises candidate character strings between two candidate endpoints; and relevant candidate character strings corresponding to the candidate endpoints are subjected to natural language processing, and target character strings corresponding to the selected candidate character strings are determined to serve as selected texts corresponding to the text selection operation. Compared with the prior art, the selected texts on the touch terminal are subjected to natural language processing to determine the target character strings corresponding to the selected texts, the text selection bias caused by small operation regions or touch errors on touch terminals are overcome, and accordingly, the selection accuracy is improved, the user operation efficiency is improved, and the user operation experience is improved.

The disclosed embodiments provide a system for evaluating a performance of a mobile application. During operation, the system obtains a first set of data associated with adopters of a new version of a mobile application in a partial rollout of the new version and a second set of data associated with non-adopters of the new version in the partial rollout. Next, the system applies a statistical model to the first and second sets of data to select a subset of the non-adopters as potential adopters of the new version. The system then reduces a bias in a quasi-experimental design associated with the mobile application by using the first set of data and a third set of data associated with the potential adopters to estimate an average treatment effect (ATE) between the new version and an older version of the mobile application.

A method and system of matching an investor's objectives for portfolio investment return and risk with an assessment of a range of expected returns and risks that are likely to be generated by investment portfolios consisting at least in part of alternative asset classes that involves, for example, selecting available historical data for a plurality of alternative asset classes, unsmoothing the historical data based at least in part on historical data for traditional asset classes related to the respective alternative asset classes, and correcting the historical data for the alternative asset classes for an impact of survivorship and selection biases. A forecast of an expected return and risk is computed for each of the alternative asset classes, based at least in part on the unsmoothed and corrected historical data for the alternative asset classes, and at least one of the alternative asset classes that has an expected return and risk that corresponds substantially to the investor's objectives for portfolio investment return and risk is identified for inclusion in the investment portfolio.

Selection bias in providing user reviews for digital goods markets is compensated for by identifying users likely to provide positive reviews and actively soliciting reviews from these users. User activity data with respect to a digital good is gathered and used to identify enthusiast reviewers. Enthusiast users within the set of users are identified based on a comparison of user activity data with at least one criterion. Review solicitations are generated and provided to enthusiast users. Review solicitations are then sent to the identified users to encourage them to submit reviews for digital goods.

New computational technologies generating systematic investment portfolios by coordinating forecasting algorithms contributed by researchers are provided. Work on challenges is efficiently facilitated by the algorithmic developer's sandbox (“ADS”). Second, the algorithm selection system performs a batch of tests that selects the best developed algorithms, updates the list of open challenges and translates those scientific forecasts into financial predictions. The algorithm controls for the probability of backtest overfitting and selection bias, thus providing for a practical solution to a major flaw in computational research involving multiple testing. Third, the incubation system verifies the reliability of those selected algorithms. Fourth, the portfolio management system uses the selected algorithms to execute investment recommendations. A dynamically optimal portfolio trajectory is determined by a quantum computing solution to combinatorial optimization representation of the capital allocation problem. Fifth, the crowdsourcing of algorithmic investments controls the workflow and interfaces between all of the hereinabove introduced components.

High-throughput methods for mapping integration sites resulting from one or more integrations, such as infection by a retrovirus, are disclosed. The disclosed methods require no selection for specific phenotypes such as antibiotic resistance, and thereby may avoid selection bias. Moreover, the linker-based amplification is simple and rapid, and by using a frequently cutting restriction enzyme, the amplicons are small, which significantly decreases possible amplification and cloning biases.

A method and system of matching an investor's objectives for portfolio investment return and risk with an assessment of a range of expected returns and risks that are likely to be generated by investment portfolios consisting at least in part of alternative asset classes that involves, for example, selecting available historical data for a plurality of alternative asset classes, unsmoothing the historical data based at least in part on historical data for traditional asset classes related to the respective alternative asset classes, and correcting the historical data for the alternative asset classes for an impact of survivorship and selection biases. A forecast of an expected return and risk is computed for each of the alternative asset classes, based at least in part on the unsmoothed and corrected historical data for the alternative asset classes, and at least one of the alternative asset classes that has an expected return and risk that corresponds substantially to the investor's objectives for portfolio investment return and risk is identified for inclusion in the investment portfolio.

A device for correcting injector characteristics includes: an injector-characteristic detection unit that outputs a selection signal by detecting drive characteristics of injectors in response to a selection control signal; a selection control unit that outputs the selection control signal for selecting a factor for deviation correction to the injector-characteristic detection unit; a selection confirmation unit that confirms a variation value of the factor corresponding to the selection signal; a deviation correction unit that calculates a deviation compensation value corresponding to the injector characteristics in response to an output signal of the selection confirmation unit, and thus output a correction signal; a control unit that generates a correction clock in response to the correction signal; and an output drive unit that controls driving of the injectors in response to the correction clock.

The invention relates to a software code search oriented query statement regenerating method. The method includes the following steps: performing pretreatment on codes and comments in a software code library and query statements input by a user; extracting compound words in the software code library, and defining two or more keywords divided by the compound words as a heterogeneous relationship; defining synonyms as a homogeneous relationship; searching homogeneous keywords and heterogeneous keywords of all the keywords in the query statements, and visualizing the homogeneous keywords and the heterogeneous keywords so as to allow the user to select more suitable keywords which are the final result. The method overcomes the defects of low accuracy, selection deviation and complex relationships of the conventional methods; the method can effectively search relevant information in the software code library, expand keywords included in original query statements, provide code snippets of the keywords, achieve information query and expansion, and effectively improve software understanding level of a software maintainer and the efficiency.

The invention discloses a popular content pool maintenance method and apparatus. The method comprises the steps of extracting popular contents to a popular content pool from predetermined news sources; obtaining release time corresponding to the popular contents and feedback information for the popular contents, wherein the feedback information is used for representing a user behavior of feeding back the popular contents by a user; according to the feedback information and the release time, determining scores corresponding to the popular contents respectively, wherein the scores are used for assessing attention to the popular contents; and deleting the popular contents with the scores not meeting predetermined standards from the popular content pool. By collecting the feedback informationof the user to the released popular contents and the freshness of the popular contents, current feedback information and future possibility are balanced, and the defect of selection bias in a manual selection and automatic aggregation scheme adopted in the prior art is overcome, so that the effect of effectively fusing high feedback rate and content freshness is achieved; and the method and the apparatus have the advantages of accurately releasing the popular contents and improving the user experience.

The present invention provides a method of controlling the selection of a cell in a radio communication system for increasing the capacity of the system by effectively utilizing a small cell base station and a base station apparatus for realizing it. At least one base station of a macrocell base station and a small cell base station is provided with plural antennas, and the base station selects a cell and adjusts a criterion for reselection by generating processing gain using the plural antennas by signal processing and performing a cell selection bias correcting process using the processing gain using the plural antennas.

The invention relates to a method and a system for compiling fund indexes. The method includes the following steps: acquiring a constituent fund meeting the index setting requirements; carrying out sample cleaning and error processing on the indexes of the constituent fund; acquiring monthly frequency indexes and weekly frequency indexes in the indexes after sample cleaning and error processing; releasing the monthly frequency indexes and the weekly frequency indexes; and correcting the indexes released. The indexes are selected in strict accordance with six requirements, and are cleaned based on multiple data cleaning rules. Survival bias, backfill bias and selection bias influencing the indexes are processed. The monthly frequency indexes and the weekly frequency indexes are extracted by cross-checking multiple data sources. The indexes are released in different periods of time according to the attributes of the monthly frequency indexes and the weekly frequency indexes. The indexes released are corrected. Thus, the objectivity and authenticity of data is ensured, the indexes compiled are accurate and representative, a product truly representing the market is chosen, and an accurate investment strategy can be developed.

A device for correcting injector characteristics includes: an injector-characteristic detection unit that outputs a selection signal by detecting drive characteristics of injectors in response to a selection control signal; a selection control unit that outputs the selection control signal for selecting a factor for deviation correction to the injector-characteristic detection unit; a selection confirmation unit that confirms a variation value of the factor corresponding to the selection signal; a deviation correction unit that calculates a deviation compensation value corresponding to the injector characteristics in response to an output signal of the selection confirmation unit, and thus output a correction signal; a control unit that generates a correction clock in response to the correction signal; and an output drive unit that controls driving of the injectors in response to the correction clock.

A method and system of matching an investor's objectives for portfolio investment return and risk with an assessment of a range of expected returns and risks that are likely to be generated by investment portfolios consisting at least in part of alternative asset classes that involves, for example, selecting available historical data for a plurality of alternative asset classes, unsmoothing the historical data based at least in part on historical data for traditional asset classes related to the respective alternative asset classes, and correcting the historical data for the alternative asset classes for an impact of survivorship and selection biases. A forecast of an expected return and risk is computed for each of the alternative asset classes, based at least in part on the unsmoothed and corrected historical data for the alternative asset classes, and at least one of the alternative asset classes that has an expected return and risk that corresponds substantially to the investor's objectives for portfolio investment return and risk is identified for inclusion in the investment portfolio.

The invention discloses a method and a system for positioning a wireless sensor network. The method comprises the following steps that: after a first distance between a node to be positioned and an anchor node is calculated, a preset quantity of anchor node sub-combinations are constructed; after a second distance between the anchor node and the node to be positioned is calculated, the anchor node sub-combination which has a smallest deviation sum and the deviation sum in an allowed range is selected as a destination anchor node sub-combination; and the final position coordinates of the node to be positioned are calculated through the destination anchor node sub-combination. By the method and the system provided by the invention, the anchor node sub-combination with the smallest deviation can be selected, and measurement data with large errors can be effectively eliminated, so the positioning accuracy of the node to be positioned can be effectively improved.

The invention discloses a method for predicting a time sequence intervention effect. The method comprises the following steps: acquiring a data set; constructing a deep learning model; enabling two depth representations formed by the same input in the two recurrent neural network units to form a pair, and applying orthogonalization constraints to the pair; defining a total loss function, and minimizing the total loss function to realize training optimization of the model; training on the training set by adopting different hyper-parameter combinations, verifying the performance on the verification set, selecting the optimal hyper-parameter combination, and testing the deep learning model trained by adopting the optimal hyper-parameter on the test set to obtain a predicted intervention effect. According to the prediction method provided by the invention, related representations of intervention measures and intervention effects are respectively learned by utilizing two recurrent neural network units through a multi-task learning strategy, hybrid factors causing selection bias errors are eliminated through orthogonalization constraints, and finally, decoupling representations on a time sequence are obtained through training and are used for predicting the intervention effects of the time sequence.

The invention discloses a car holding and use behavior analysis method based on a sample selection model and further analyzes a role of the built residence place and employment place environment. The method comprises steps that 1, data is acquired, and the data mainly comprises large-scale resident travel data, social economic attributes and the built residence place and work place environment data; 2, a car holding selection equation is determined according to the sample selection model; 3, a car use result equation is constructed for a car holding group acquired through selection in the step 2, and the result equation is added into selection deviation items; 4, a model is constructed in statistic software, and the built environment is set to be the built residence place and employment place environment; and 5, practical significance and reasons implied in each parameter of the model result are analyzed. The method is advantaged in that endophytism and dependence relationships existing in car holding and use are considered, a combined model based on the sample selection model is utilized, sample selection deviation is eliminated, and influence of the built residence place and employment place environment on car dependence is studied.

The embodiment of the invention provides a decorrelation clustering method and device under data selection deviation, and the decorrelation clustering method comprises the steps: obtaining a pluralityof images with deviation, and enabling the images to serve as a sample set; based on the sample set, combining and optimizing a post-weighting clustering algorithm and a decorrelation regular term toobtain an optimal post-weighting clustering algorithm, obtaining the optimal post-weighting clustering algorithm by calculating the post-weighting clustering algorithm for multiple times, and obtaining the post-weighting clustering algorithm by using the weight of each sample obtained by learning the decorrelation regular term to weight the clustering algorithm, wherein the weight of each sampleis obtained by learning the weight of each sample of each image in the sample set by using a decorrelation regular term; and obtaining an optimal weighted clustering algorithm when a current clustering center and a cluster contained in the current weighted clustering algorithm are not the first clustering center and the cluster and the difference between the current clustering center and the cluster and the last clustering center and the cluster is smaller than a threshold value, so that the clustering center and the cluster, not affected by deviation, of an image are determined.

The embodiment of the invention provides an agricultural Internet of Things data multi-view projection clustering reconstruction method and system, and the method comprises the steps of obtaining theperception data sent by the agricultural Internet of Things, and generating a perception matrix according to the perception data; projecting the perception matrix on a time plane, a space plane and aparameter plane to obtain a time plane projection matrix, a space plane projection matrix and a parameter plane projection matrix; respectively clustering the time plane projection matrix, the space plane projection matrix and the parameter plane projection matrix through a preset clustering algorithm to respectively obtain a deviation value of the time plane projection matrix, a deviation value of the space plane projection matrix and a deviation value of the parameter plane projection matrix; and selecting data in the projection matrix with the minimum deviation value as reconstructed data to carry out agricultural Internet of Things data reconstruction. According to the method provided by the embodiment of the invention, through a multi-view projection technology, the execution complexity of a clustering algorithm is reduced, and the efficient and accurate reconstruction of multi-dimensional complex agricultural WSN data is realized.

The invention discloses a peritoneal dialysis treatment effect prediction system based on variational inference and deep learning. The system comprises an information acquisition module, a calculationprocessing module, an auxiliary recommendation module and a self-learning module; the calculation processing module adopts a prediction model based on variational inference and deep learning; the process comprises the steps of obtaining a retrospective experiment data set; performing deriving to obtain a variation lower bound, and converting the maximum likelihood function into the maximum variation lower bound; constructing a corresponding model, and taking maximization of a variation lower bound as an optimization target; selecting an optimal hyper-parameter combination by using hyper-parameter search; and testing a model which adopts optimal hyper-parameter training on the test set. The model can predict the expected treatment effect difference of automatic peritoneal dialysis and manual peritoneal dialysis for an individual under the condition of giving individual characteristics, and decouples hidden variables through a variational inference method, so that the influence of selection bias errors on prediction is reduced, and more accurate prediction performance is obtained. And a decision maker can be better assisted in selecting a treatment mode.

A non-parallel storage life test evaluation method based on the selection of the acceleration factor feasible region, the steps are as follows: Step 1: Calculate the acceleration factor and the comprehensive failure time; Step 2: Discuss the acceleration factor feasible region; Step 3: Use the best linear Partial estimation method for parameter selection; Step 4: Calculate the point estimation and interval estimation of the logarithmic reliable life. The present invention aims at the phenomenon of missing storage data in the life evaluation of non-parallel storage products. On the basis of the service life obeying the Weibull distribution, the storage data and the test data are integrated through the acceleration factor, and the data are fully utilized for calculation, ensuring the accuracy of the life model parameter estimation. Accuracy, its algorithm has lower requirements on the initial value of the parameters, the algorithm iteration is fast and simple, and the operability is strong.

The invention discloses a popular content pool maintenance method and apparatus. The method comprises the steps of extracting popular contents to a popular content pool from predetermined news sources; obtaining release time corresponding to the popular contents and feedback information for the popular contents, wherein the feedback information is used for representing a user behavior of feeding back the popular contents by a user; according to the feedback information and the release time, determining scores corresponding to the popular contents respectively, wherein the scores are used for assessing attention to the popular contents; and deleting the popular contents with the scores not meeting predetermined standards from the popular content pool. By collecting the feedback informationof the user to the released popular contents and the freshness of the popular contents, current feedback information and future possibility are balanced, and the defect of selection bias in a manual selection and automatic aggregation scheme adopted in the prior art is overcome, so that the effect of effectively fusing high feedback rate and content freshness is achieved; and the method and the apparatus have the advantages of accurately releasing the popular contents and improving the user experience.

In the present invention, when trace image data is acquired while using a probe (16) to scan an area on a sample (10) in a forward direction and retrace image data is acquired while scanning the samearea in the reverse direction, a deviation information storage unit (32) stores deviation detected by a deviation detection unit (20). This deviation is an indication of the difference between the distance between the probe and the sample and a target value for the distance at a given point in time. An image data selection unit (31) compares the deviation during forward scanning and the deviationduring reverse scanning for each measurement point, selects the image data acquired during scanning that has the smallest deviation, and stores the same to a storage area (33c) of an image data storage unit (33) as selected image data. An image formation unit (30) creates a sample surface shape image on the basis of the selected image data and displays the same on a display unit (5). As a result,even if there is a steep recess or protrusion on a sample (10) and there is a tracking delay or disturbance in the feedback control for the distance between the probe and the sample, it is possible todisplay an accurate image close to the actual shape of the surface of the sample.

The invention relates to the technical field of data automatic intelligent processing and discrimination system, and discloses a trans-esophageal echocardiography-guided left aurcle closure device selection system based on artificial intelligence. The trans-esophageal echocardiography-guided left aurcle closure device selection system based on artificial intelligence includes a data collection total system, a cloud platform data storage system, a data modeling and analysis system, and a data display system, wherein the data collection total system is in communication connection to the cloud platform data storage system; the cloud platform data storage system is connected to the data modeling and analysis system; and the data display system is connected to the data modeling and analysis system. The trans-esophageal echocardiography-guided left aurcle closure device selection system based on artificial intelligence selects the type and size of a left aurcle closure device according to the discriminating result data, so that the software can automatically select a closure device most suitable for the patient according to the left aurcle image data provided by esophageal echocardiography, thereby reducing improper selection of instruments, caused by subjective selection bias of the operator, reducing the radiation intake and patient contrast intake of the operator, and ultimately increasing the success rate of surgery.

The present invention relates to the field of detection, in particular to the use of a method for purifying, isolating and analyzing atypical circulating tumor cells and the use of atypical circulating tumor cells. The method for purifying, isolating and analyzing the atypical circulating tumor cells, as well as the atypical circulating tumor cells, can be used for preparing kits diagnosing cancer, assessing cancer prognosis, monitoring cancer, or assessing effectiveness of cancer treatment. The invention has the advantages of high throughput, high purity and high recovery rate, and no selection bias, successful purification, isolation and analysis of atypical circulating tumor cells. Through the analysis of atypical circulating tumor cells, cancer risk, cancer prognosis and effectivenesscan be assessed, and the sensitivity and accuracy of detection can be improved.