Past Prices Research Articles

Artificial Intelligence: Can Seemingly Collusive Outcomes Be Avoided?

Strategic decisions are increasingly delegated to algorithms. We extend previous results of the algorithmic collusion literature to the context of dynamic optimization with imperfect monitoring by analyzing a setting where a limited number of agents use simple and independent machine-learning algorithms to buy and sell a storable good. No specific instruction is given to them, only that their objective is to maximize profits based solely on past market prices and payoffs. With an original application to battery operations, we observe that the algorithms learn quickly to reach seemingly collusive decisions, despite the absence of any formal communication between them. Building on the findings of the existing literature on algorithmic collusion, we show that seeming collusion could originate in imperfect exploration rather than excessive algorithmic sophistication. We then show that a regulator may succeed in disciplining the market to produce socially desirable outcomes by enforcing decentralized learning or with adequate intervention during the learning process. This paper was accepted by Gabriel Weintraub, revenue management and market analytics. Supplemental Material: The data files are available at https://doi.org/10.1287/mnsc.2022.4623 .

The Lead-Lag Relationship Between International Food Prices, Freight Rates, and Trinidad and Tobago's Food Inflation: A Support Vector Regression Analysis

The objective of this study is to empirically investigate: i) the extent in which shipping costs can cause and predict the food inflation in Trinidad and Tobago (T and T); ii) the extent in which international food prices can drive food inflation in T and T; and iii) the impact of the Russia-Ukraine conflict on food prices in T and T. As a fourth sub-objective, this study seeks to provide policy recommendations to address the food price inflation in T and T. A structural model is sought as it can use information about the past prices of the international food prices and shipping rates to forecast T and T's food price inflation. this study used monthly data on T and T retail price index food subcategory, the FAO Food Price Index (FFPI), and the Freightos Baltic Index over the January 2015 to November 2022 period. the Support Vector Regression (SVR) model was applied and used the FFPI and Freightos Baltic Index as the leading indicators to forecast T and T's food price inflation. this study found that the FAO Food Prices does and the Freightos Baltic Index does have an impact on food prices in T and T. Secondly, based on the FAO index and the Freightos Baltic Index, T and T's food price inflation should be 14.91% lower than its present value. third, in a Markov Switching regression, the war dummy was only significant in the second regime, statistically suggesting that the Russia-Ukraine conflict only had a short term effect on T and T's food prices. this contributed to the literature as it introduced a new test to investigate predictive causality. Policy measures to increase the agriculture output and address the food price inflation in T and T include adopting sustainable farming practices such as integrated pest management T and The sustainable water management, and encouraging aquaponics.

Understanding the Price Behaviour of selected banks from banking sector in Bombay Stock Exchange

Over the past two decades, the globalisation trend has led to a significant increase in financial markets, particularly the stock market. Stock markets play an important role as an indicator of the economy's overall performance. Analyzing past price movements is how Technical Analysis predicts future price movements. It encourages the mobilization of funds, provides equity capital to the banking sector, and offers adequate returns for investors. The Indian banking sector has undergone significant changes in recent years, with banks diversifying from traditional deposit-taking and lending activities to more complex activities such as investment banking, insurance, and wealth management. Therefore, understanding the price behavior of banks listed on the Bombay Stock Exchange (BSE) is crucial for investors, policymakers, and researchers. By using stock market indicators, investors can observe significant market movements. The main objective of this paper is to analyze a top 5 bank within the banking sector and forecast its stock price behavior. Additionally, the paper aims to test the effectiveness of technical indicators such as RSI and MACD in predicting the stock's behavior. The study focuses on analyzing the stock price trends of selected banks over a period of five years (2017-2022).

Comprehensive Study of Machine learning Algorithms for Stock Market Prediction during COVID-19

The stock market has always drawn investors' attention. Stock trend forecasting tools are in high demand because they facilitate the direct acquisition of profits. The more precise the results, the greater the likelihood of generating a greater profit. Statistically, only 3% of the Indian population invests in stocks, and at the time of COVID-19, that number was even lower, as the stock market was not based on general patterns and equations but rather on the emotional quotient of the people. Such a circumstance increased the stock market's vulnerability. The pandemic factor, such as the case of COVID-19, influences the stock market's trends. Technical analysis of market trends is a technique for interpreting past and present prices to forecast likely future prices. Several deep learning and machine learning algorithms are utilized to generate stock market forecasts, wherein LSTM and ARIMA models have been proven to produce reasonably accurate results. The prior works focused on individual models and their components to provide forecasts. Thus, this paper aims to compare the two well-established models and provide investors with models that work well with data and have appropriate parameter values. The LSTM and ARIMA models are presented because they provide appropriate results using technical analysis of the data set, and the results are to be compared. The closing values from the historic stock prices of the top three Indian hospitality industries were used as the data set. The results show that the individual models work well when the data matches the model and the appropriate parameter is used.

거래량과 가격정보에 따른 모멘텀과 주식수익률의 관계에 관한 연구

According to behavioral finance, investors make investment decisions by recognizing past information as useful information. Investment decisions based on past information cause an overreaction to stock prices. According to previous studies on the Korean stock market, momentum based on trading volume and price information has a significant positive influence on stock returns. In this study, the Korean stock market is divided into the securities stock market and the KOSDAQ market to investigate which market the existing results are due to. Furthermore, an empirical analysis is conducted on the irrational behavior of investors based on transaction volume and price information. The research method uses univariate portfolio analysis and multivariate portfolio analysis and divides the decile portfolio and the conditional 5th quintile portfolio based on the momentum substitution considering the transaction volume. It is checked whether the momentum based on past transaction volume and price information lasts for the next six months. As a result, in the securities market, it is confirmed that the momentum according to transaction volume and price information has a significant positive effect on stock returns, and these results are maintained even after considering firm characteristics. On the other hand, it is stated that momentum according to transaction volume and price information does not occur in the KOSDAQ market. It shows that the existing results are due to the securities stock market rather than the KOSDAQ market.

In Art We Trust

Whereas trust is the cornerstone of any market’s functioning, it is of particular importance in markets that are unregulated, illiquid, and opaque, such as the art market. This study examines the role of authenticity, as captured by provenance information in auction catalogs, on the probability of auctioned oil paintings, watercolors, and prints being sold; their price formation; and returns. Auction catalogs include four authenticity dimensions: pedigree (ownership “blockchain,” descendance information; type of past owners, such as renowned collectors; and past sales records), exhibition history (e.g., in famous museums or galleries), literature coverage (e.g., in catalogues raisonnés or authoritative press), and certification (e.g., artist’s physical testimonial, expert opinions). We find that trust, proxied by provenance information, increases the probability of a work being sold by up to 4%, leads to hammer price premiums up to 54%, and increases annualized returns by 5%–16%. To address potential endogeneity problems between the provision of provenance and past prices/price expectations, we perform quasi-natural experiments in difference-in-differences settings on auction houses’ provenance policy changes following authenticity litigation and on a contamination effect of the discovery of fakes and forgeries on the oeuvre of forged artists. We also test transactions less affected by past prices, such as estate sales following the death of a collector. The findings on the relation between provenance and prices are robust to artist reputation, artistic style, auction house reputation, art market liquidity, and artist career timing. This paper was accepted by Tomasz Piskorski, finance. Funding: Y. Li gratefully acknowledges supports from the National Natural Science Foundation of China [Grant 72204257] and the Area Studies Fund of Renmin University of China [Grant AS2022002]. This study is partially funded by the Tilburg Alumni Fund. Supplemental Material: The online appendix and data are available at https://doi.org/10.1287/mnsc.2022.4633 .

Exploring the application of machine learning and python model lstm in predicting Tesla stock price

Machine learning involves many fields such as probability theory, statistics, and algorithm complexity theory. Using existing algorithms and models to make judgments or predictions without clear instructions, relying on existing data, patterns, and reasoning. It is a subset of Artificial Intelligence that includes statistical techniques and knowledge to deal with a dataset with vast amounts of data. It can be applied to our daily life, such as investment decisions of stocks, options futures and other derivatives. In this article, the author will predict price performance by using a just-in-time LSTM model. The goal of the mission is to use past historical price data to make accurate predictions about Tesla's stock price over the future. Due to the characteristics of Tesla's case of high-volatility stocks, this problem is extremely challenging and has become a popular research topic in stock price forecasting. From a machine learning perspective, this is a good, critical but challenging prediction problem. That's because Tesla is already a leader in electric vehicle (EV) production. The linear regression model and LSTM (long and short-term model) model are respectively used to forecast Tesla's stock price. The RMSE value is used to evaluate the accuracy of the two models and the results are obtained: the accuracy of the LSTM model prediction is much higher than that of a linear regression model.

Technical Analysis of Digital Currencies and the Most Important Indicators for Trading

Technical analysis studies chart patterns, using indicators based on past price action data to predict future trends. Technical analysis involves mathematical calculations based on an asset’s price or volume, with the results used to predict future prices and can indicate a market trend or warnings that the trend is about to reverse. No single indicator will detect a market reversal or confirm a market trend. The best strategy is to use a combination of indicators, or better, choose only a few of the available tools and combine them without too much chart cluttering, which might lead to more confusion than clarity. The tools are available, and traders will personalize their experience choosing the ones that better fit their trading style. No specific indicators will offer better results than others; it’s mainly down to whether the trader feels more comfortable using one indicator over another. This article will cover some of the most popular and utilized crypto research and investment tools available in the crypto space. However, this is by no means an exhaustive guide of resources available. With more experience, a trader can keep learning and understand the countless possibilities of other technical analysis tools.

The role of the past long-run oil price changes in stock market

This paper examines the ability of the past long-run changes in oil price to predict the stock returns in the U.S. market. We find this long-lag model performs much better than the one-lag model. The past long-run changes in oil price contain useful information about future real stock returns and excess returns over a Treasury bill rate. This variable alone can capture more than 1% variations of next horizon (month) excess returns, and the predictive power are increasingly strong for long-horizon stock return. These findings are robust when considering other popular predictors into the model, these results are also maintained when considering various subsamples. For out-of-sample examination, the results of McCraken’s (2007) Ros2 and Clark and West’s (2007) MSPE-adjusted statistic explore that this variable contains useful information of future stock returns. More interestingly, the past long-run oil price changes also perform strong predictive power on excess returns for non-US countries.

Making Money: a LR+ARIMA and Multi-Index Based Investment Model

In order to achieve the goal of maximizing returns, market traders tend to buy and sell assets with high volatility. Now, in this paper, traders will buy and sell only two assets, namely gold, and bitcoin, based on their requirements, using past daily prices to determine the trading prices to determine whether they should buy, continue to hold, or sell the assets in their portfolio on that date. Determine a portfolio consisting of cash, gold, and bitcoin that is based on an investment model for a five-year trading period starting at $1,000 on September 11, 2016, and ending on September 10, 2021. During this period, they will follow a trading strategy when appropriate. (The gold market is only open on weekdays and the bitcoin market is open daily with commission fees for each trade).

Predictability of stock returns using neural networks: Elusive in the long term

The rapidly growing neural network literature continually reports successful stock price forecasting results. Many of these studies use relatively short evaluation periods, spanning only a couple of years. In this paper, sustainability of the neural network forecast quality over the long term is analysed. Feedforward and recurrent networks are used to predict the direction of monthly stock price movements, with past price data as predictors. The analysis is conducted on the NYSE stocks over the 1971–2015 period and all the evaluations are performed out-of-sample. Statistically significant directional predictability for selected assets is found. However, the trading simulations reveal that directional predictability does not guarantee trading performance better than the benchmark buy-and-hold strategy. The opportunities for investors to use the tested models for profit appear to be episodic and periodically enhanced e.g. in periods of recession.

Machine Learning and the Stock Market

AbstractPractitioners allocate substantial resources to technical analysis whereas academic theories of market efficiency rule out technical trading profitability. We study this long-standing puzzle by applying a diverse set of machine learning algorithms. The results show that an investor can find profitable technical trading rules using past prices, and that this out-of-sample profitability decreases through time, showing that markets have become more efficient over time. In addition, we find that the evolutionary genetic algorithm’s attitude in not shying away from erroneous predictions gives it an edge in building profitable strategies compared to the strict loss-minimization-focused machine learning algorithms.

Time Series Analysis and Forecasting of Gold Price using ARIMA and LSTM Model

Abstract: Gold is an age-old method of investing money. It is tangible and can be passed on from one generation to another. It is the one form of investment that many people consider very safe to employ, to keep their money safe and easily multiplicative. Investment guides and consultants usually read charts to predict the future price of this commodity. In this research paper, we are making use of Machine Learning models to predict the price of gold based on past prices. The dataset consists of the opening, closing, highest and lowest prices of gold through 7 years, daily. This is a comparative study that highlights the best model, between a classical statistical model ARIMA and a recurrent neural network model LSTM

Beliefs asymmetry and price stability in a cobweb model

It is known that a cobweb model with heterogeneous expectations and evolutionary selection on beliefs is able to generate irregular and complex price dynamics. I show in this work that a necessary condition for such dynamics to emerge is the asymmetry in the set of available predictors over which evolutionary forces operate. If predictors are symmetric around the rational one, heterogeneous expectations aggregate up to rationality, and the unique rational expectations price is stable. If predictors are asymmetrically distributed, instead, the outcome depends on the side of the asymmetry: a prevalence of predictors that overshoot price changes leads to market stability, irrespective on any dynamics on beliefs, while a predominance of predictors anchored to past prices can be destabilizing. It is only in this last case that an evolutionary selection mechanism can give rise to complex dynamics in prices. As the distribution of beliefs depends on the set of heuristics allowed to compete in the market, the design of this set is shown to be crucial for the possible outcomes of the model.

Heuristic pricing rules not requiring knowledge of the price response function

Heuristic rules are appropriate, if a decision maker wants to set the price of a new product or of a product, whose past price variation is low, and budget limitations prevent the use of marketing experiments or customer surveys. Whereas such rules are not guaranteed to provide the optimal price, generated profits should be as close as possible to their optimal values. We investigate eleven pricing rules that do not require that a decision maker knows the price response function and its parameters. We consider monopolistic market situations, in which sales depend on the price of the respective product only. A Monte Carlo simulation that is more comprehensive than extant attempts found in the literature, serves to evaluate these rules. The best performing rules either hold price changes between periods constant or make them dependent on the previous absolute price difference. These rules also outperform purely random price setting, which we use as benchmark. On the other hand, rules based on arc elasticities or on a loglinear approximation to sales and prices, turn out to be even worse than random price setting. In the conclusion, we discuss how heuristic pricing rules may be extended to deal with product line pricing, additional marketing variables (e.g., advertising, sales promotion, and sales force) and a duopolistic market situation.

Replenishment Policy with Limited Price Information

In this paper, we seek to provide real-time replenishment policy for a capacitated company that faces an exogenous demand and has to procure and store sufficient quantities of certain items with uncertain procurement prices. The objective is to minimize the total cost of the procuring cost and inventory holding cost without stock out. Motivated by the fact that a large variability in the price process often makes forecasting difficult, in contrast to the traditional approaches of the stochastic inventory theory, we study this problem from the perspective of competitive analysis who is free of any price distribution assumption. We propose an online replenishment model, which generalizes several previous online models. We give a real-time replenishment policy whose decisions are made based entirely on the past and present prices and the current inventory level. We prove that our policy achieves superior performance, and outperforms all of the previous theoretical results. Furthermore, we present the numerical studies to show that our replenishment policy achieves an even better empirical performance.

High-Low Promotion Policies for Peak-End Demand Models

In-store promotions are a highly effective marketing tool that can have a significant impact on revenue. In this research, we study the question of dynamic promotion planning in the face of Bounded-Memory Peak-End demand models. In order to determine promotion strategies, we establish that a High-Low pricing policy is optimal under diagonal dominance conditions (so that the current period price dominates both past period price effects and competitive product price effects on the demand), as well as conditions on the price dispersion. We show that finding the optimal High-Low dynamic promotion policy is NP-hard in the strong sense. Nevertheless, for the special case of promotion planning for a single item, we propose a compact Dynamic Programming (DP) approach that can find the optimal promotion plan that follows a High-Low policy in polynomial time. When the diagonal dominance conditions do not hold, and, hence, a High-Low policy is not necessarily optimal, we show that the optimal High-Low policy that is found by our proposed DP can find a provably near-optimal solution. Using the proposed DP as a subroutine, for the case of multiple items, we propose a Polynomial-Time-Approximation Scheme (PTAS) that can find a solution that can capture at least [Formula: see text] of the optimal revenue and runs in time that is exponential only in [Formula: see text]. Finally, we test our approach on data from large retailers and demonstrate an average of [Formula: see text] increase in revenue relative to the retailer’s current practices. This paper was accepted by Chung Piaw Teo, optimization. Funding: Financial support from the Oracle Corporation [External Research Office grant] and the National Science Foundation [Grant CMMI-1162034] is gratefully acknowledged.

Research on the Investment Strategy of Quantitative Trading Based on Random Forest and Genetic Algorithm

Gold and Bitcoin are two common investments in the market. Investors profit from the capital market through specific trading strategies. Based on previous price data, this paper established a prediction-decision model, which provides investors with an optimal trading strategy and effectively improves the return on investment. To begin with, we use the past daily price of gold and bitcoin to calculate the Bollinger band, DMA, and MACD characteristics. We used the Random Forest algorithm to predict the next day's price and established a loop to predict all price data. Finally, we got the prediction data with the goodness of fit ( ) as high as 0.98. We intuitively see that the prediction effect is pretty good from the predictive value-real value-line chart. Then, we established the transaction decision-making objective model, used the Genetic Algorithm to select the daily transaction amount, obtained the optimal transaction strategy, and then established a loop to obtain all the optimal strategies. The final total assets were 243,424.76, and the annual return rate reached 6085.62%.

Role of credit and expectations in house price dynamics

Using quarterly data for 18 advanced economies from 1991–2020 we estimate a fixed-effects quantile dynamic panel model and provide evidence for non-linearity in the effects of credit, interest rates and expectations on house price growth. We find that excessive credit buildup negatively impacts house price growth during housing busts, whereas low interest rates and expectations, as measured by past house price growth is associated with housing price booms. Using local projections, we also find evidence that these effects persist over multiple quarters.

Deep Reinforcement Learning-Based Demand Response for Smart Facilities Energy Management

This work proposes a novel deep reinforcement learning (DRL)-based demand response algorithm for smart facilities energy management to minimize electricity costs while maintaining a satisfaction index. Specifically, to accommodate the characteristics of the decision-making problem, long short-term memory (LSTM) units are adopted to extract discriminative features from past electricity price sequences and fed into fully connected multi-layer perceptrons (MLPs) with the measured energy and time information, then a deep Q-network is developed to approximate the optimal policy. After that, an experimental setup is constructed to investigate the effectiveness of the proposed DRL-based demand response algorithm to bridge the gap between theoretical studies and practical implementations. Numerical results demonstrate that the proposed algorithm can handle energy management well for multiple smart facilities. Moreover, the proposed algorithm outperforms the model predictive control (MPC) strategy and uncontrolled solution and is close to the theoretical optimal control method.