Algorithm for Detecting Acoustic Traces of Neutrino Decay

Askold Belyakov — Fri, 01 Mar 2024 15:52:48 +0700

In the articles "Doppler Effect and Acoustic Trails of Neutrinos" and "Neutrinos above the Earth's surface", published in the journal "LA MULTIAPP" on February 25, 2022 and July 12, 2023, I was forced to touch on the astrophysical topic when, in the flow of data obtained in the process Monitoring acoustic noise with a recorder with a Nyquist frequency of 5 kHz, at a depth of 1000 m in the earth's crust, events began to be frequently observed, the shapes and frequencies of which are not typical for seismology and geophysics, and had not previously been observed by anyone. The shape of some events was very similar to the traces of neutrino decay, such as the bipolar pulse and the “diamond”, recorded by acoustic detectors in seawater. Subsequently, when recording high-frequency acoustic processes on the Earth's surface with a recorder with a Nyquist frequency of 50 kHz, events were discovered that were even more similar to traces of neutrino decay. At the same time, a “strange” association of some events of unknown origin with the field extrema of the power electrical network of a large city was discovered. To avoid erroneous conclusions, it is necessary to understand the origin of this “strangeness”.

Bitcoin Price Prediction Model Development Using Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM)

Jonathan Cahyadi, Amalia Zahra — Thu, 07 Mar 2024 11:52:00 +0700

Cryptocurrency is a virtual currency that can be used as a financial or economic standard, foreign currency reserve, and as a means of payment in some countries. The value that goes up and down every time is not easy to predict using logic. This is a problem for investors, besides that investors lack knowledge about the direction of crypto money movement. In addition, there is no system that can predict the price of Bitcoin, so this can cause investors to take the wrong steps in transactions and can cause losses. To avoid this risk, a system is needed that can predict bitcoin prices using data mining techniques, namely forecasting, the algorithms used are CNN and LSTM. The data used is Bitcoin closing price data from January 1, 2017, to April 26, 2023. The data is divided into 80% training data and 20% testing data. The prediction results are evaluated using MAPE which gets a MAPE value of 0.037 or 3.7% in the CNN algorithm, while the LSTM algorithm gets a value of 0.065 or 6.5%. The MAPE results of the two algorithms are in the MAPE range <10%, so it can be said that the ability of the forecasting model is very good so that it can be used as a reference to determine the prediction of bitcoin prices in the next few periods.

Journal La Multiapp

Algorithm for Detecting Acoustic Traces of Neutrino Decay

Bitcoin Price Prediction Model Development Using Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM)