Chapter I: Classical Analysis provides a good overview of the data for AstroPower, and reveals patterns that will be explored with detail in later sections. |
| Prev. Co. | Next Co. | |
| ||
| This explains the analysis charts for AstroPower: The first chart contains the original date, in this case, the monthly average closing prices for APWR over history. Blue marks the de-seasonalized trend, which is a least squares linear regression applied to the APWR prices after seasonal variations have been filtered out. The first step of the classical analysis determines seasonal indexes. The decomposition above bases the seasonality on the ordinary 12 month calendar from one January to the next, but it is also possible to extract interesting results from the 24 or 48 month political election based on November elections, as we shall see shortly. Here the seasonal component is plotted in orange. Because it is based on the seasonality of the entire time series, it is regular from one year to the next. Once the global seasonal is known, it is possible to subtract its influence from the original input to produce so called deseasonalized data. The trend line in the top chart comes about from this processes. Further refinement removes the trend from the deseasonalized data. What remains is the unfiltered cyclical component. Broadly speaking, the refined cyclical data represents the effect of the general business cycle in addition to the private business cycle of AstroPower. Later, we will use a method to separate the two cycles, but on this chart, they are compounded. The purple chart is the Irregular Component. This classical name is not entirely appropriate, since it often reveals obvious regular patterns. Because it represents the variations that have not been explained by the refinement process up to this step, it could be called the Un-explained Component. One of the more interesting series to be derived in this manner is the red trace on the bottom chart. The residue that is left when all components other than Seasonal and Un-Explained are filtered away, shows how the strongly regular Seasonal effects actually change from one year to the next. When a seasonal pattern becomes well know, the market may anticipate, causing the date of the seasonal peak to occur earlier. This component show how that anticipation moves over time, unlike the static pattern in the second chart. While the full history analysis is of enduring interest, it may also be misleading. An issue may have the same name for 20 or more years, but it may be a very different company, and economic fundamentals have surely changed. For this reason, we want to look at patterns established in more recent times. |
| Prev. Co. | Next Co. | |
 | ||
Refined Stock Trend Analysis for APWR : |