A very reliable credit estimation algorithm
Posted on : 15-05-2010 | By : admin | In : sideline business, small business, stock market, taxes, transactions
Tags: bad debt, car loans, compare credit, currency trading, forex, funds, home equity, portfolio
0
Some intraday data series do not contain a buy/sell identifier for trades. As most estimators of transaction costs require this information, an algorithm is necessary to classify trades as buy or sell. When bid/ask quotes (or best buy/sell limit order prices) are available, a natural method is to compare the trade price with the quotes prevailing at the time of the trade. Trades at or above the ask (or best sell limit price) will be classified as buys, trades at or below the bid (or best buy limit price) will be classified as sells. This algorithm will leave trades within the quotes unclassified. Trades within the quotes may be either ‘crosses’ (trades negotiated outside the central market place) or trades where a floor broker or the order book improved on the price of the specialist’s quote, as often happens on hybrid markets like the NYSE. Lee and Ready (1991) show that this algorithm is very reliable and classifies most trades in the same way as the classification based on a comparison with bid and ask quotes.



