Given a data set D containing millions of data points and a data consumer who is willing to pay for $X to train a machine learning (ML) model over D, how should we distribute this $X to each data point to reflect its "value"? In this paper, we define the "relative value of data" via the Shapley value, as it uniquely possesses properties with appealing real-world interpretations, such as fairness, rationality and decentralizability. For general, bounded utility functions, the Shapley value is known to be challenging to compute: to get Shapley values for all N data points, it requires O(2N) model evaluations for exact computation and O(N log N) for (ϵ, δ)-approximation.In this paper, we focus on one popular family of ML models relying on K-nearest neighbors (KNN). The most surprising result is that for unweighted KNN classifiers and regressors, the Shapley value of all N data points can be computed, exactly, in O(N log N) time - an exponential improvement on computational complexity! Moreover, for (ϵ, δ)-approximation, we are able to develop an algorithm based on Locality Sensitive Hashing (LSH) with only sublinear complexity O(Nh(ϵ, K) log N) when ϵ is not too small and K is not too large. We empirically evaluate our algorithms on up to 10 million data points and even our exact algorithm is up to three orders of magnitude faster than the baseline approximation algorithm. The LSH-based approximation algorithm can accelerate the value calculation process even further.We then extend our algorithm to other scenarios such as (1) weighed KNN classifiers, (2) different data points are clustered by different data curators, and (3) there are data analysts providing computation who also requires proper valuation. Some of these extensions, although also being improved exponentially, are less practical for exact computation (e.g., O(NK) complexity for weigthed KNN). We thus propose an Monte Carlo approximation algorithm, which is O(N(log N)2/(log K)2) times more efficient than the baseline approximation algorithm.

/pdf/efficient-task-specific-data-valuation-for-nearest-neighbor-26cs0j2drv.pdf

Efficient task-specific data valuation for nearest neighbor algorithms

Diagnosis of schizophrenia (SZ) is traditionally performed through patient’s interviews by a skilled psychiatrist. This process is time-consuming, burdensome, subject to error and bias. Hence the aim of this study is to develop an automatic SZ identification scheme using electroencephalogram (EEG) signals that can eradicate the aforementioned problems and support clinicians and researchers. This study introduces a methodology design involving empirical mode decomposition (EMD) technique for diagnosis of SZ from EEG signals to perfectly handle the behavior of non-stationary and nonlinear EEG signals. In this study, each EEG signal is decomposed into intrinsic mode functions (IMFs) by the EMD algorithm and then twenty-two statistical characteristics/features are calculated from these IMFs. Among them, five features are selected as significant feature applying Kruskal Wallis test. The performance of the obtained feature set is tested through several renowned classifierson a SZ EEG dataset. Among the considered classifiers, theensemble bagged tree performed as the best classifier producing 93.21% correct classification rate for SZ, with an overall accuracy of 89.59% for IMF 2. These results indicate that EEG signals discriminate SZ patients from healthy control (HC) subjects efficiently and have the potential to become a tool for the psychiatrist to support the positive diagnosis of SZ.

A Computerized Method for Automatic Detection of Schizophrenia Using EEG Signals

The rapid development of information technology and the fast growth of Internet have facilitated an explosion of information which has accentuated the information overload problem. Recommender systems have emerged in response to this problem and helped users to find their interesting contents. With increasingly complicated social context, how to fulfill personalized needs better has become a new trend in personalized recommendation service studies. In order to alleviate the sparsity problem of recommender systems meanwhile increase their accuracy and diversity in complex contexts, we propose a novel recommendation method based on social network using matrix factorization technique. In this method, we cluster users and consider a variety of complex factors. The simulation results on two benchmark data sets and a real data set show that our method achieves superior performance to existing methods.

A novel recommendation method based on social network using matrix factorization technique

Machine learning in medical applications: A review of state-of-the-art methods

In the field of Artificial Intelligence Machine learning provides the automatic systems which learn and improve itself from experience without being explicitly programmed. In this research work a movie recommender system is built using the K-Means Clustering and K-Nearest Neighbor algorithms. The movielens dataset is taken from kaggle. The system is implemented in python programming language. The proposed work deals with the introduction of various concepts related to machine learning and recommendation system. In this work, various tools and techniques have been used to build recommender systems. Various algorithms such as K-Means Clustering, KNN, Collaborative Filtering, Content-Based Filtering have been described in detail. Further, after studying different types of machine learning algorithms, there is a clear picture of where to apply which algorithm in different areas of industries such as recommender systems, e-commerce, etc. Then there is an illustration of how implementations and working of the proposed system are used for the implementation of the movie recommender system. Various building blocks of the proposed system such as Architecture, Process Flow, Pseudo Code, Implementation and Working of the System is described in detail. Finally, in this work for different cluster values, different values of Root Mean Squared Error are obtained. In this proposed work as the no of clusters decreases, the value of RMSE also decreases. The best value of RMSE obtained is 1.081648. The results given by the proposed system are better than the existing technique on the basis of RMSE value.

Movie Recommender System Using K-Means Clustering AND K-Nearest Neighbor

https://cyberleninka.org/article/n/1071663.pdf

D.A. Adeniyi

Papers

Automated web usage data mining and recommendation system using K-Nearest Neighbor (KNN) classification method