this is for holding javascript data
Yoav Ram renamed Mandates forecasting.md to Seats distribution forecasting.md
about 9 years ago
Commit id: baa62c4b70843024a8bb4145d98e41dd1df7eb58
deletions | additions
diff --git a/Seats distribution forecasting.md b/Seats distribution forecasting.md
new file mode 100644
index 0000000..9575c4a
--- /dev/null
+++ b/Seats distribution forecasting.md
...
## Seats distribution forecasting
We developed several mandates calculation algorithms. Here, we will only describe our latest algorithm with some variations, as we believe that this is our best attempt at seats forecasting.
The basic problem is how to standardize our voters sample, which, although consisting of over 7,000 voters, can be biased due to several factors such as age, socio-economical status, and party activist propaganda. Our current approach to control the sample biases was designed together with Omri Dor.
We started asking users for their 2013 elections choices on February 13th 2015. We use this information, together with the 2013 elections [official results](http://www.votes-19.gov.il/nationalresults).
First, we take only the latest vote for each device id, both from the 2013 and the 2015 datasets. Next, we calculate a counts matrix $C$ with rows for 2015 parties, columns for 2013 parties and values for the number of votes in each row-column combination. Thus, $C_{i,j}$ is the number of voters that voted for party $j$ in 2013 and will vote for party $i$ in 2015.
Next, we use the counts matrix $C$ to estimate the transition matrix $M$ in which $M_{i,j}$ is the probability that an individual who voted for party $j$ in 2013 will vote for party $i$ in 2015.
We now generate the 2013 results vector $v$ from the results data, removing counts of parties for which we have no information and illegal or discarded votes.
We multiply the transition matrix by the results vector to get the forecast vector $f = C \cdot v$. The forecast vector $f$ now describes our prediction of the number of votes each party will get in the 2015 elections.
To get a forecast of the number of seats for each party we then process the votes forecast vector $f$ using the [Bader-Offer method](https://www.knesset.gov.il/lexicon/eng/seats_eng.htm), also known as the [Hagenbach-Bischoff system](http://en.wikipedia.org/wiki/Hagenbach-Bischoff_system). In our version of the Bader-Offer method we disregarded surplus vote agreements.
As another layer of bias correction, we experimented with fixing of number of votes received by four major demographies to the number of votes in 2013. These demographies are:
1. The arab sector, represented by Hadash, Balad & Raam-Taal in 2013 and by the Arab Unified List in 2015.
2. The Ashkenazi-Orhodox sector, represented by Yahadut Ha'Tora both in 2013 and in 2015.
3. The Sfaradi-Orthodox secotr, represented by Shash and Am Shalem in 2013 and by Shas and Yachad in 2015. Because Yachad merged with Ozma La'Am for the 2015 elections, we includied Ozma La'Am in the respective 2013 votes.
4. The liberal, pro-cannabis legalisation party, Ale Yarok.
Fixing the number of voters of the first three demographies over a two year span can be justified due the the relatively constant number of seats the respective parties received in the previous three elections and by the sectoriality of these parties. As for fixing the number of votes of Ale Yarok, this was deemed as neccessary because supporters of this party are known to be very active online, thus generating biases in online surveys and polls. For example, the number of "Likes" Ale Yarok has in Facebook is 85,709, compared with 27,205 Ha'Likud, the ruling party, has.
